AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
March 1991
CAN THE NONSTRESS TEST PREDICT CONGENITAL SEPSIS? Ron Gonen, M.D., Arne Ohlsson, M.D., M.Sc, Dan Farine, M.D., and John E. Milligan, M.D.
ABSTRACT
Preterm premature rupture of the membranes constitutes one of the most common complications of pregnancy. A recent survey among hospitals in North America revealed that even though there is no consensus regarding the management of patients with preterm premature rupture of the membranes, over 90% of the responders favored the conservative approach.1 However, the conservative approach carries an increased risk of fetal or neonatal infection. Indeed, some management protocols include amniocentesis in patients with preterm premature rupture of the membranes to detect not only the fetus with a mature lung profile, but also the fetus most likely to develop sepsis.23 Others depend on nonspecific clinical signs of chorioamnionitis, such as fever, leukocytosis, foul-smelling lochia, tender uterus, fetal and maternal tachycardia, and an elevated C-reactive protein to detect overt or impending infections. Recently, it has been suggested that an abnormal nonstress test alone or in conjunction with other biophysical parameters is a good predictor of congenital sepsis.4"7 The present retrospective study was undertaken to investigate further the role of the nonstress test in predicting congenital sepsis in preg-
Downloaded by: Georgetown University Medical Center. Copyrighted material.
The possible role of the nonstress test (NST) in the prediction of congenital sepsis in pregnancies complicated by prolonged premature rupture of the membranes (PROM) was investigated in a retrospective controlled study. Thirteen NSTs performed within 24 hours of the delivery of infants with congenital sepsis were coded, blinded, and mixed with an equal number of blinded NSTs derived from randomly selected patients matched for gestational age with PROM but without neonatal or maternal infection. Eleven of 13 infants with congenital sepsis had a reactive NST, and two had a nonreactive test. Among the controls, there were also 11 reactive and two nonreactive NSTs. The sensitivity and specificity of the NST in predicting congenital sepsis in patients with PROM were 15.4% and 84.6%, respectively. The positive predictive value and the negative predictive value were 50%. Our results suggest that the NST is not a useful tool for prediction of congenital sepsis.
nancies complicated by preterm premature rupture of the membranes. MATERIALS AND METHODS
The study was conducted at the University of Toronto Perinatal Complex. Initially, we identified all neonates born between January 1, 1984, and December 31, 1987, with congenital sepsis. Congenital sepsis was defined as the combination of positive blood cultures at birth and clinical symptoms of sepsis. We then searched our nonstress test files to identify which mothers of septic neonates had a nonstress test within 24 hours of delivery. The last nonstress test of each of the study patients was coded, blinded, and mixed with an equal number of blinded nonstress tests derived from randomly selected patients matched for gestational age with premature rupture of the membranes but without neonatal or maternal infection. For the purpose of this report a reactive nonstress test was defined as one with two or more fetal heart rate accelerations of at least 15 beats/min in amplitude and at least 15 sec-
University of Toronto, Regional Perinatal Complex, Women's College Hospital and Mount Sinai Hospital, Toronto, Ontario, Canada Requests for reprint: Dr. Milligan, Perinatal Unit, Women's College Hospital, 76 Grenville St., Toronto, Ontario, M5S 1B2 Canada Present address: Ron Gonen, M.D., 60 Harofe St., Haifa 34367, Iarael Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
91
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
RESULTS
Thirteen infant-mother pairs qualified for the study, and their last nonstress tests were evaluated together with 13 controls. Table 1 illustrates patients' characteristics and clinical data. There were no differences between the two groups in terms of maternal age, gestational age, number of days with ruptured membranes, the rate of caesarean sections, infants' weight, and Apgar scores. Six of the study patients had clinical signs of amnionitis. Twenty-two nonstress tests were reactive and four were nonreactive. Eleven of 13 infants with congenital sepsis had a reactive nonstress test within 24 Table 1. Patients' Characteristics and Clinical Data Congenital Sepsis (n = 13) Gestational age (wk) Maternal age PROM* duration (days) Clinical amnionitis Caesarean section Weight (gm) Apgar 1 minute Apgar 5 minute
Controls (n = 13)
29.2 ± 2.3 27.8 8 ± 6.2 6 (46%) 6 (46%) 1380 ± 468
29.1 ± 2.2
8 (62%) 1350 ± 382
4 7.2
5 7.5
29
13 ± 12.4 0
"PROM: Premature rupture of the membranes.
hours of delivery and two had a nonreactive test. Among the controls, there were also 11 reactive and two nonreactive nonstress tests. The sensitivity and specificity of the nonstress test in predicting congenital sepsis in patients with premature rupture of the membranes were 15.4% and 84.6%, respectively. The positive predictive value and the negative predictive value were both 50%. The clinical data of the 13 infants with congenital sepsis are displayed in Table 2. The two most common organisms isolated on blood cultures (and in most cases from additional sites as well) were Escherichia coli, isolated from five infants, and group B Streptococcus, isolated from four infants. The remaining four infants were infected by one of the following: Streptococcuspneumoniae, Candida albicans, Moraxella osloensis, and Serratia marcescense. There
were four neonatal deaths, all in the study group, two associated with E. coli sepsis, one with group B Streptococcus, and one with C. albicans. The two nonreactive nonstress tests were not associated with neonatal death and the infecting organisms were E. coli and group B Streptococcus.
DISCUSSION
The optimal method for monitoring patients with preterm premature rupture of the membranes remains to be determined. Clinical symptoms and signs of infection and white cell count are insensitive indicators of intra-amniotic infection. Amniocentesis is highly accurate in predicting chorioamnionitis, but much less accurate in predicting fetal sepsis.2'3-6 Recently, Vintzileos et al4 suggested that the nonstress test is a very sensitive and relatively specific predictor of infectious outcome in patients with premature rupture of the membranes. In 32 pregnancies with either amnionitis, possible neonatal sepsis, or neonatal sepsis, the nonstress test was nonreactive in 25 and reactive in seven. The sensitivity and specificity were 78.1% and 86.3%, respectively. In nine
Table 2. Clinical Data of Infants with Congenital Sepsis Patient 1 2
3 4
5 6 7 8
9 10
1 I 12 13 92
Gestation at Birth
PROM* Days
Clinical Amnionitis
Mode of Delivery
25 26 27 28 28 28 30 30 31 31 32 32 32
5 14 5 4
Yes
Breech C.S. C.S. C.S.
9 4
3 3 12 20 20 2
3
No Yes Yes No Yes No Yes No Yes No No No
SVD
Forceps Forceps SVD SVD
C.S. C.S. C.S. SVD
Weight 630 810 870
1140 1080 1330 1580 1510 1925 1260 2345 1730 1730
Organism Grown GBS
E. coli S. pneumonae Serratia E. coli E. coli Candida albicans GBS
E. coli E. coli Moraxella osloensis GBS GBS
Neonatal Outcome
NST*
Alive Alive Alive Alive Alive Dead Dead Dead Dead Alive Alive Alive Alive
NR R R R R R R R R NR R R R
*PROM: Premature rupture of the membranes; NST: nonstress test; SVD: spontaneous vaginal delivery; R: reactive; N: nonreactive; GBS: Group B: hemolytic Streptococcus.
Downloaded by: Georgetown University Medical Center. Copyrighted material.
onds duration in a 20-minute period. A nonstress test was defined as being nonreactive if it did not fulfill these criteria. The clinical diagnosis of amnionitis was made in the presence of two or more of the following criteria: maternal temperature higher than 37.8°C, maternal tachycardia greater than 120 beats/min, leukocytosis more than 20,000/mm3 (in the absence of prior corticosteroid administration), fetal tachycardia more than 160 beats/min, uterine tenderness, and foulsmelling amniotic fluid.
March 1991
NST AS PREDICTOR OF CONGENITAL SEPSIS/Gonen, Ohlsson, Farine, Milligan
on this test in the management of patients with premature rupture of the membranes. We conclude that, based on our results, the nonstress test is not a useful tool for early prediction of congenital sepsis. Larger and prospective studies are needed to investigate further the role of the nonstress test and various biophysical parameters in the diagnosis of fetal infection rather than chorioamnionitis.
REFERENCES
1. Capeless EL, Mead PM: Preterm premature rupture of membranes: Lack of a national consensus. Am J Obstet Gynecol 157:11-12, 1987 2. Garite TJ, Freeman RK, Linzey M, Braly P: The use of amniocentesis in patients with premature rupture of the membranes. Obstet Gynecol 54:226-230, 1979 3. Cotton DB, Hill LM, Strassner HT, Platt L, Ledger WJ: The use of amniocentesis in preterm gestation with ruptured membranes. Obstet Gynecol 63:38-43, 1984 4. Vintzileos AM, Campbell WA, Nochimson DJ, Weinbaum PJ: The use of nonstress tests in patients with premature rupture of the membranes. Am J Obstet Gynecol 155: 149-153, 1986 5. Ventzileos AM, Campbell WA, Nochimson DJ, Connolly ME, Fuenfer MM, Hoehn GJ: The fetal biophysical profile in patients with premature rupture of the membranes. An early predictor of fetal infection. Am J Obstet Gynecol 152:510-516, 1985 6. Vintzileos AM, Campbell WA, Nochimson DJ, Weinbaum PJ, Mirochnick MH, Escoto DT: Fetal biophysical profile versus amniocentesis in predicting infection in preterm premature rupture of the membranes. Obstet Gynecol 68:488-494, 1986 7. Vintzileos AM, Bors-Koefoed R, Pelegano JF, et al: The use of fetal biophysical profile improves pregnancy outcome in premature rupture of the membranes. Am J Obstet Gynecol 157:236-240, 1987
Downloaded by: Georgetown University Medical Center. Copyrighted material.
cases of neonatal sepsis, nonreactive nonstress test was encountered in seven patients. Subsequently, the same group5 assessed the predictive value of the biophysical profile for the development of infection: If a biophysical profile performed within 24 hours of delivery scored equal or greater than 8 of 12 the infection rate was 2.7%. If, on the other hand, the score was less than 8, it was associated with an infection rate of 93.7%. The first manifestations of fetal infection in that study were nonreactive nonstress tests and loss of fetal breathing. In fact, 15 of 16 infected patients had nonreactive nonstress tests. By including a low biophysical profile (in the presence of nonreactive nonstress tests and absence of fetal breathing) among the indications for delivery of patients with premature rupture of the membranes, Vintzileos et al7 have shown a lower infectious outcome when compared with controls whose management were based on clinical parameters with or without amniocentesis. A low biophysical profile was also found to be a more sensitive tool than amniocentesis in predicting neonatal sepsis rather than amnionitis.6 Again, nonreactive nonstress tests was present in four of five patients. In the present study, a nonstress test performed within 24 hours of delivery was reactive in 11 of 13 cases of proven congenital sepsis. We have no explanation as to why our results differ so much from those of the group from the University of Connecticut,4^7 apart from differences in methodology. It is not the low positive predictive value of the nonreactive nonstress test that is our main concern, because such patients will be further evaluated. It is the low negative predictive value and the poor sensitivity of the nonstress test in predicting fetal sepsis that may give false reassurance if one were to depend
93
March 1991
CAN THE NONSTRESS TEST PREDICT CONGENITAL SEPSIS? Ron Gonen, M.D., Arne Ohlsson, M.D., M.Sc, Dan Farine, M.D., and John E. Milligan, M.D.
ABSTRACT
Preterm premature rupture of the membranes constitutes one of the most common complications of pregnancy. A recent survey among hospitals in North America revealed that even though there is no consensus regarding the management of patients with preterm premature rupture of the membranes, over 90% of the responders favored the conservative approach.1 However, the conservative approach carries an increased risk of fetal or neonatal infection. Indeed, some management protocols include amniocentesis in patients with preterm premature rupture of the membranes to detect not only the fetus with a mature lung profile, but also the fetus most likely to develop sepsis.23 Others depend on nonspecific clinical signs of chorioamnionitis, such as fever, leukocytosis, foul-smelling lochia, tender uterus, fetal and maternal tachycardia, and an elevated C-reactive protein to detect overt or impending infections. Recently, it has been suggested that an abnormal nonstress test alone or in conjunction with other biophysical parameters is a good predictor of congenital sepsis.4"7 The present retrospective study was undertaken to investigate further the role of the nonstress test in predicting congenital sepsis in preg-
Downloaded by: Georgetown University Medical Center. Copyrighted material.
The possible role of the nonstress test (NST) in the prediction of congenital sepsis in pregnancies complicated by prolonged premature rupture of the membranes (PROM) was investigated in a retrospective controlled study. Thirteen NSTs performed within 24 hours of the delivery of infants with congenital sepsis were coded, blinded, and mixed with an equal number of blinded NSTs derived from randomly selected patients matched for gestational age with PROM but without neonatal or maternal infection. Eleven of 13 infants with congenital sepsis had a reactive NST, and two had a nonreactive test. Among the controls, there were also 11 reactive and two nonreactive NSTs. The sensitivity and specificity of the NST in predicting congenital sepsis in patients with PROM were 15.4% and 84.6%, respectively. The positive predictive value and the negative predictive value were 50%. Our results suggest that the NST is not a useful tool for prediction of congenital sepsis.
nancies complicated by preterm premature rupture of the membranes. MATERIALS AND METHODS
The study was conducted at the University of Toronto Perinatal Complex. Initially, we identified all neonates born between January 1, 1984, and December 31, 1987, with congenital sepsis. Congenital sepsis was defined as the combination of positive blood cultures at birth and clinical symptoms of sepsis. We then searched our nonstress test files to identify which mothers of septic neonates had a nonstress test within 24 hours of delivery. The last nonstress test of each of the study patients was coded, blinded, and mixed with an equal number of blinded nonstress tests derived from randomly selected patients matched for gestational age with premature rupture of the membranes but without neonatal or maternal infection. For the purpose of this report a reactive nonstress test was defined as one with two or more fetal heart rate accelerations of at least 15 beats/min in amplitude and at least 15 sec-
University of Toronto, Regional Perinatal Complex, Women's College Hospital and Mount Sinai Hospital, Toronto, Ontario, Canada Requests for reprint: Dr. Milligan, Perinatal Unit, Women's College Hospital, 76 Grenville St., Toronto, Ontario, M5S 1B2 Canada Present address: Ron Gonen, M.D., 60 Harofe St., Haifa 34367, Iarael Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
91
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
RESULTS
Thirteen infant-mother pairs qualified for the study, and their last nonstress tests were evaluated together with 13 controls. Table 1 illustrates patients' characteristics and clinical data. There were no differences between the two groups in terms of maternal age, gestational age, number of days with ruptured membranes, the rate of caesarean sections, infants' weight, and Apgar scores. Six of the study patients had clinical signs of amnionitis. Twenty-two nonstress tests were reactive and four were nonreactive. Eleven of 13 infants with congenital sepsis had a reactive nonstress test within 24 Table 1. Patients' Characteristics and Clinical Data Congenital Sepsis (n = 13) Gestational age (wk) Maternal age PROM* duration (days) Clinical amnionitis Caesarean section Weight (gm) Apgar 1 minute Apgar 5 minute
Controls (n = 13)
29.2 ± 2.3 27.8 8 ± 6.2 6 (46%) 6 (46%) 1380 ± 468
29.1 ± 2.2
8 (62%) 1350 ± 382
4 7.2
5 7.5
29
13 ± 12.4 0
"PROM: Premature rupture of the membranes.
hours of delivery and two had a nonreactive test. Among the controls, there were also 11 reactive and two nonreactive nonstress tests. The sensitivity and specificity of the nonstress test in predicting congenital sepsis in patients with premature rupture of the membranes were 15.4% and 84.6%, respectively. The positive predictive value and the negative predictive value were both 50%. The clinical data of the 13 infants with congenital sepsis are displayed in Table 2. The two most common organisms isolated on blood cultures (and in most cases from additional sites as well) were Escherichia coli, isolated from five infants, and group B Streptococcus, isolated from four infants. The remaining four infants were infected by one of the following: Streptococcuspneumoniae, Candida albicans, Moraxella osloensis, and Serratia marcescense. There
were four neonatal deaths, all in the study group, two associated with E. coli sepsis, one with group B Streptococcus, and one with C. albicans. The two nonreactive nonstress tests were not associated with neonatal death and the infecting organisms were E. coli and group B Streptococcus.
DISCUSSION
The optimal method for monitoring patients with preterm premature rupture of the membranes remains to be determined. Clinical symptoms and signs of infection and white cell count are insensitive indicators of intra-amniotic infection. Amniocentesis is highly accurate in predicting chorioamnionitis, but much less accurate in predicting fetal sepsis.2'3-6 Recently, Vintzileos et al4 suggested that the nonstress test is a very sensitive and relatively specific predictor of infectious outcome in patients with premature rupture of the membranes. In 32 pregnancies with either amnionitis, possible neonatal sepsis, or neonatal sepsis, the nonstress test was nonreactive in 25 and reactive in seven. The sensitivity and specificity were 78.1% and 86.3%, respectively. In nine
Table 2. Clinical Data of Infants with Congenital Sepsis Patient 1 2
3 4
5 6 7 8
9 10
1 I 12 13 92
Gestation at Birth
PROM* Days
Clinical Amnionitis
Mode of Delivery
25 26 27 28 28 28 30 30 31 31 32 32 32
5 14 5 4
Yes
Breech C.S. C.S. C.S.
9 4
3 3 12 20 20 2
3
No Yes Yes No Yes No Yes No Yes No No No
SVD
Forceps Forceps SVD SVD
C.S. C.S. C.S. SVD
Weight 630 810 870
1140 1080 1330 1580 1510 1925 1260 2345 1730 1730
Organism Grown GBS
E. coli S. pneumonae Serratia E. coli E. coli Candida albicans GBS
E. coli E. coli Moraxella osloensis GBS GBS
Neonatal Outcome
NST*
Alive Alive Alive Alive Alive Dead Dead Dead Dead Alive Alive Alive Alive
NR R R R R R R R R NR R R R
*PROM: Premature rupture of the membranes; NST: nonstress test; SVD: spontaneous vaginal delivery; R: reactive; N: nonreactive; GBS: Group B: hemolytic Streptococcus.
Downloaded by: Georgetown University Medical Center. Copyrighted material.
onds duration in a 20-minute period. A nonstress test was defined as being nonreactive if it did not fulfill these criteria. The clinical diagnosis of amnionitis was made in the presence of two or more of the following criteria: maternal temperature higher than 37.8°C, maternal tachycardia greater than 120 beats/min, leukocytosis more than 20,000/mm3 (in the absence of prior corticosteroid administration), fetal tachycardia more than 160 beats/min, uterine tenderness, and foulsmelling amniotic fluid.
March 1991
NST AS PREDICTOR OF CONGENITAL SEPSIS/Gonen, Ohlsson, Farine, Milligan
on this test in the management of patients with premature rupture of the membranes. We conclude that, based on our results, the nonstress test is not a useful tool for early prediction of congenital sepsis. Larger and prospective studies are needed to investigate further the role of the nonstress test and various biophysical parameters in the diagnosis of fetal infection rather than chorioamnionitis.
REFERENCES
1. Capeless EL, Mead PM: Preterm premature rupture of membranes: Lack of a national consensus. Am J Obstet Gynecol 157:11-12, 1987 2. Garite TJ, Freeman RK, Linzey M, Braly P: The use of amniocentesis in patients with premature rupture of the membranes. Obstet Gynecol 54:226-230, 1979 3. Cotton DB, Hill LM, Strassner HT, Platt L, Ledger WJ: The use of amniocentesis in preterm gestation with ruptured membranes. Obstet Gynecol 63:38-43, 1984 4. Vintzileos AM, Campbell WA, Nochimson DJ, Weinbaum PJ: The use of nonstress tests in patients with premature rupture of the membranes. Am J Obstet Gynecol 155: 149-153, 1986 5. Ventzileos AM, Campbell WA, Nochimson DJ, Connolly ME, Fuenfer MM, Hoehn GJ: The fetal biophysical profile in patients with premature rupture of the membranes. An early predictor of fetal infection. Am J Obstet Gynecol 152:510-516, 1985 6. Vintzileos AM, Campbell WA, Nochimson DJ, Weinbaum PJ, Mirochnick MH, Escoto DT: Fetal biophysical profile versus amniocentesis in predicting infection in preterm premature rupture of the membranes. Obstet Gynecol 68:488-494, 1986 7. Vintzileos AM, Bors-Koefoed R, Pelegano JF, et al: The use of fetal biophysical profile improves pregnancy outcome in premature rupture of the membranes. Am J Obstet Gynecol 157:236-240, 1987
Downloaded by: Georgetown University Medical Center. Copyrighted material.
cases of neonatal sepsis, nonreactive nonstress test was encountered in seven patients. Subsequently, the same group5 assessed the predictive value of the biophysical profile for the development of infection: If a biophysical profile performed within 24 hours of delivery scored equal or greater than 8 of 12 the infection rate was 2.7%. If, on the other hand, the score was less than 8, it was associated with an infection rate of 93.7%. The first manifestations of fetal infection in that study were nonreactive nonstress tests and loss of fetal breathing. In fact, 15 of 16 infected patients had nonreactive nonstress tests. By including a low biophysical profile (in the presence of nonreactive nonstress tests and absence of fetal breathing) among the indications for delivery of patients with premature rupture of the membranes, Vintzileos et al7 have shown a lower infectious outcome when compared with controls whose management were based on clinical parameters with or without amniocentesis. A low biophysical profile was also found to be a more sensitive tool than amniocentesis in predicting neonatal sepsis rather than amnionitis.6 Again, nonreactive nonstress tests was present in four of five patients. In the present study, a nonstress test performed within 24 hours of delivery was reactive in 11 of 13 cases of proven congenital sepsis. We have no explanation as to why our results differ so much from those of the group from the University of Connecticut,4^7 apart from differences in methodology. It is not the low positive predictive value of the nonreactive nonstress test that is our main concern, because such patients will be further evaluated. It is the low negative predictive value and the poor sensitivity of the nonstress test in predicting fetal sepsis that may give false reassurance if one were to depend
93
