European Journal of Obstetrics & Gynecology and Reproductive Biology, 44 (1992) 53-58 53

0 1992 Elsevier Science Publishers B.V. All rights reserved OO28-2243/92/$05.00

EUROBS 01282

The predictive value of fetal heart rate monitoring:

a retrospective analysis of 2165 high-risk pregnancies

V. Ocak, F. Demirkiran, C. Sen, U. Colgar, F. 6$er, 6. Kilavuz and Y. Uras

Department of Obstetrics and Gynecology, Division of Perinatology, Cerrahpasa Medical School, ZJniL,ersityof Istanbul, Turkey

Accepted for publication 27 September 1991

Summary

The predictive value of fetal heart-rate monitoring on fetal well-being was studied in 2165 high-risk
pregnancies. 1883 reactive nonstress test (NST) patterns and 278 nonreactive NST patterns and 4 cases
of sinusoidal pattern were obtained. Oxytocin challenge test (OCT) was applied to 263 nonreactive cases.
OCT was not applied to 15 cases out of 278 nonreactive NST cases, because of placenta previa,
abruptio placenta and previous cesarean section. There were 155 cases with negative OCT, 84 cases with
positive OCT and 24 cases with equivocal, prolonged or severe variable decelerations. Sensitivity and
specificity were for NST 50 and 88% and for OCT 60 and 67%. The positive and negative predictive
values were 11 and 98% for NST and 18 and 93% for OCT. It is concluded that the reactive nonstress
test is a reliable test for good outcome but a positive oxytocin challenge test is not a reliable test for poor
outcome. Additional procedures are necessary such as assessment of fetal growth, doppler velocity
waveforms and fetal biophysical profile to avoid unnecessary obstetric interventions and to reach good
fetal outcome.

FHR monitoring; Non-stress test; Oxytocin challenge test; Apgar score; Sensitivity; Specificity

Introduction many authors concluded that false-negative OCT

rate was excessive [6,71. We have employed fetal
Electronic fetal heart-rate monitoring is widely heart rate monitoring since 1978. FHR monitor-
used to assess fetal condition in high-risk preg- ing was carried out in 2165 high-risk pregnancies
nancies. NST and OCT have been employed in following the fixed protocol shown in Table I.
USA since 1970 [1,2]. In Europe, Kubli and Ham- The aim of the current study was to assess the
macher have started the initial studies on FHR predictive value of NST and OCT in relation to
monitoring 13-51. Many studies were published the Apgar scores at birth.
about the predictive value of NST and OCT in
relation to prognosis of fetal outcome, however Materials and Methods

The patient population included 2165 high-risk

Correspondence: Dr. V. Ocak, Department of Obstetrics and pregnancies who were managed between 1978-
Gynecology, Division of Perinatology, Cerrahpasa Medical 1988 in the Division of Perinatoiogy, Department
School, University of Istanbul, Turkey. of Obstetrics and Gynecology, Cerrahpasa Medi-
54

TABLE I ultrasonographic evaluation at admission. There

FETAL HEART RATE MONITORING PROTOCOL were 99 cases with poor obstetric history, which
included previous stillbirth, repeated miscarriage,
NST unexplained intrapartum and neonatal death and
Reactive 2 Nonreactive previous pre-eclampsia. There were 79 cases with
serious maternal systemic disorders, which in-
cluded diabetes mellitus, maternal cardiac dis-
Weekly or
OCT ease, systemic lupus erythematodes, sickle disease
twice a week
(prolonged pregnancy, and familial Mediterranean fever.
diabetes rnellitus) -Negative l+ Positive
Monitoring procedures
Until 44 weeks Estimatio I of lung The NST protocols of Rochard and Schifrin
maturation were applied to all high-risk patients [1,2]. We
Induction of labor used a Hewlett-Packard 8030 A Model car-
C/S diotocography for fetal monitoring. Fetal heart-
If failed two times
rate tests were performed after meals. During the
test the patients maintained in a semi-fowler po-
sition. If two or more FHR accelerations, exceed-
C/S
ing an amplitude of 15 beats and a duration of 15
s occurred in association with fetal movements
during a 10 min period, the nonstress test (NST)
was considered reactive. Nonreactive NST was
cal Faculty, University of Istanbul. Maternal age the failure to meet these criteria [9]. NST was
varied between 17 and 55 years. Most patients performed to all the patients once a week, except
were in the range 20-28 years. Table II shows the in prolonged pregnancies and diabetic pregnan-
indications for antepartum testing. There were cies. In these instances the test was performed
1697 prolonged pregnancies with a pregnancy du- twice a week. If a reactive pattern was found, the
ration more than 42 weeks. 267 out of 1697 patient was sent home and received an appoint-
prolonged pregnancies had ‘uncertain dates’. ment for the following week.
These cases were managed according to the same If a reactive pattern was not seen within 20
protocol as was applied for prolonged pregnan- min, we stimulated the fetus with either abdomi-
cies (Table I). The gestational age of the uncer- nal palpation or by giving a glucose containing
tain cases was at least at term on the basis of the beverage. If accelerations were not seen within 40
min, the pattern was accepted as nonreactive. If
the result was nonreactive, oxytocin was adminis-
TABLE II tered according to the protocol [8,9]. Oxytocin
Primary test indications challenge tests (OCT) were classified as negative,
positive or equivocal according to the criteria of
n % Schifrin [lo], irrespective of whether contractions
Prolonged pregnancy 1697 78.4 were spontaneous or induced. All of the traces
Poor obstetric history 99 4.6 were interpreted by the same clinician. If OCT
Rh isoimmunization 80 3.7
was negative, another appointment was given for
Maternal systemic disorders 79 3.6
Hypertension in pregnancy 72 3.3 the next week. When the OCI was positive, the
Premature labor 35 1.6 patients were delivered by cesarean section after
Others a 103 4.6 estimation of lung maturation, if pregnancy dura-
tion was less than 34 weeks. If an equivocal result
a Suspected IUGR, IUGR, advanced maternal age, uterine
malformation, consangineous marriages, decreased fetal was obtained, NST and OCT were repeated the
movement, oligohydramnios. next day. In the presence of prolonged or severe
 55

variable decelerations during NST or OCT, the TABLE IV

patients were delivered as with positive OCT. The results of OCT in 263 cases L’
Only the latest test before delivery was accepted
as a valid test for evaluation. Fetal outcome was Results of OCT No. of cases %

evaluated by means of the 5-minute Apgar score. Negative 155 59.0

The fetal monitoring protocol is shown in Table Positive 84 31.9
Equivocal 10 3.8
I. Patients with a reactive NST were followed
Prolonged variable
biweekly until 44 weeks of gestation when labor decelaration 14 5.3
was induced. We proceeded with cesarean sec-
Total 263 100.0
tion when the induction failed twice.
a In the 15 cases of placentae previa, abruptic placentae,
Statistical analysis premature labor, previous cesarean section, OCT test were
Sensitivity, specificity and Chi-square analysis not performed.

were used for statistical evaluation.

Apcar score less than 7 in 278 nonreactive NST

Results
(11.2%). As shown in Table VI, ten cases had an
Apgar score less than 7 in 155 negative OCTs
The results of the NSTs are shown in Table
(6.5%) and 15 cases in 84 positive OCTs (17.9%).
III. 1883 of 2165 NSTs were reactive and 278
nonreactive. Four cases demonstrated a sinu-
Discussion
soidal pattern, all of them in association with Rh
isoimmunization. The results of the OCTs carried
Electronic fetal heart rate monitoring tests
out after a nonreactive NST are shown in Table
such as NST and OCT are widely used to deter-
IV. OCT was carried out in 263 of 278 cases with
mine fetal well-being. Although NST is not reli-
nonreactive NST. The remaining 15 patients were
able before 32 weeks of gestation, OCT can be
cases such as placenta previa, signs of abruptio
used after 28 weeks [11,12]. For this reason we
placenta or a previous cesarean section in which
included the patients with preterm labor after 28
OCT was considered to be contraindicated. One
weeks of gestation in the study. As shown in
hundred and fifty-five of 263 OCTs were nega-
tive, and 84 of them were positive. There were 10 Table II, most patients had prolonged pregnan-
cases with equivocal and 14 cases with prolonged cies (78.4%). In prolonged pregnancies perinatal
or severe variable decelerations that were man- morbidity and mortality rates still remain high.
aged the same as in the case of a positive OCT. For this reason we have studied especially the
These 24 cases were not included in the evalua- patients with prolonged pregnancies and uncer-
tion of the Apgar scores. The Apgar scores of the tain dates. Ranking obstetric history was second
NST cases are shown in Table V. There were 31 after prolonged pregnancy in the high risk factors
cases with an Apcar score less than 7 in 1883 (Table II). This group consisted of patients with a
reactive NST cases (l-6%) and 31 cases with an history of stillbirth, intrapartum fetal loss, re-

TABLE III TABLE V

The results of NST in 2165 cases The relationship between NST and 5-min Apgar score

Results of NST No. of cases % Apgar score Reactive NST Nonreactive NST

Reactive NST 1883 87.9 <7 31 (1.65%) 31 (11.15%)

Nonreactive NST 278 12.9 a-7 1852 (98.35%) 247 (88.85%)
Sinusoidal pattern 4 0.2
Total 1883 278
Total 2165 100.0
Chi-square = 78.3; P < 0.001.
56

TABLE VI distress. One third of the neonates had low Ap-

The relationship between OCT and 5-min Apgar score gar scores [Ml.
As shown in Table V, 31 cases of the 1883 with
Apgar score Negative OCT Positive OCT a reactive NST pattern were delivered with an
<7 10 (6.45%) 15 (17.85%) Apgar score less than 7, and 1852 cases were
>7 145 (93.55%) 69 (82.15%) delivered with 7 or higher. In the nonreactive
Total 155 84 NST group, 31 cases were delivered with an Ap-
gar score less than 7 and 247 were delivered with
Chi-square: 7.566; P < 0.05.
7 or higher. While the ratio of cases with an
Apgar score less than 7 was 1.65% in the reactive
NST group, it was 11.15% in the nonreactive
peated abortions, unexplained intrapartum and NST group. The difference between the two
neonatal death and previous preeclampsia. The groups is statistically significant (P < 0.001). The
group ‘others’ mentioned in Table II included sensitivity of NST is 50% and the positive predic-
suspected IUGR, IUGR, advanced maternal age, tive value 11%. The specificity of NST is 88%
uterine malformation, consanguineous marriages, and negative predictive value 98%. It can be
decreased fetal movements and oligohydramnios. concluded that NST is a reliable test for good
A number of studies have been published in outcome but is not for poor outcome (Table VII).
the literature, which compare NST versus OCT In the negative OCT group of 155 cases, 10
as the primary test to evaluate fetal well-being cases were delivered with an Apgar score less
[8,13]. There are also several studies that show than 7, and 145 cases were delivered with 7 or
low perinatal mortality and morbidity rates if higher. In the positive OCT group of 84 cases, 15
pregnancies have been monitored applying the cases were delivered with less than 7 and 69 cases
fetal biophysical profile [14,17]. We preferred were delivered with 7 or higher. While the ratio
NST because it is a quick, easy and noninvasive of cases with an Apgar score less than 7 was
test. If the NST was reactive whatever the risk 6.45% in the negative OCT group, it was 17.85%
factor, the test was performed again in one week in the positive OCT group. The difference be-
except in prolonged pregnancies and diabetic pa- tween two OCT groups is statistically significant
tients to whom the test was applied twice a week. (P < 0.05). The sensitivity of OCT is 60% and
Sixty-two percent of our patients were delivered positive predictive value is 18%. The specificity of
vaginally and 33% by cesarean section. Our sec- OCT is 67% and negative predictive value is
tion rate may appear to be high, but all patients 93%. Again it can be concluded that OCT is a
had serious high risk factors or a previous poor reliable test for good outcome but is not for a
obstetric outcome. Fleischer et al., showed that lower Apgar score (Table VII).
one third of the patients with prolonged preg- When we compare the results from the NST
nancy delivered by cesarean section due to fetal and OCT, NST has a higher specificity (88%) and

TABLE VII
The statistical analysis

Test Sens. Spec. False ( + 1 False ( - 1 Post. Pre. Neg. Pre. Accuracy
rate rate value value rate
NST 50 88 88 1.65 11 98 88
OCT 60 67 82 6.0 18 93 61

x2: 2.61 ~2: 16.77

P > 0.05 P < 0.001
 57

negative predictive value (98.3%) than OCT. The between NST and OCT for prediction of low
difference in negative predictive value between Apgar scores and that both of them are not a
NST and OCT is statistically significant (P < reliable test for poor outcome since NST and
0.01). It means that NST is a more reliable test OCT test have a rather low positive predictive
for good outcome (98.3%) than OCT. In some value.
studies also other authors have defined that reac- Fox published that false-positive ratio, false-
tive NST was a reliable test for good outcome, negative ratio, sensitivity and specificity for OCT
Evertson et al. found that this ratio was 99% [19]. were 76, 7, 27 and 92%, respectively [26,27].
Manning et al. concluded that 96% of the NST Huddleston concluded that the false-positive rate,
cases with a reactive pattern are going to have a false-negative rate, sensitivity and specificity were
baby with an Apgar score more than 7 [20]. It was 97, 1, 50 and 81%, respectively [28]. In our study
also shown that perinatal mortality and having a we found that the false-positive rate, false-nega-
baby with a lower than seven Apgar score rates tive rate, sensitivity and specificity were 82, 6, 60
were found only in association with reactive NST and 67 for OCT test, respectively (Table VII).
pattern [9,21]. However, if deaths due to congeni- In conclusion, the general conception is that a
tal anomalies, birth trauma and neonatal sepsis negative test provides assurance that the fetus is
have been excluded, it has been found that the likely to survive and will be born in a good
surveillance rate was 99.9% for the cases of reac- condition. But the positive test result with regard
tive NST pattern [22]. Indeed, the great majority to the prediction of poor outcome is controver-
of fetus with nonreactive NST pattern have good sial. Finally, we believe that a reactive nonstress
outcome [9]. test is a reliable test for good outcome but a
Manning et al. evaluated the predictive value positive oxytocin challenge test is not a reliable
of NST according to Apgar score: The false-posi- test for poor outcome. Additional procedures are
tive rate was 2.39, positive predictive value was necessary such as assessment fetal growth,
13.1, sensitivity was 57.1, specificity was 84, nega- doppler velocity waveforms and fetal biophysical
tive predictive value was 98 and accuracy was 83.6 profile to avoid unnecessary obstetric interven-
[17]. In the present study the false-positive rate tions and to reach good fetal outcome.
was 1.65, positive predictive value was 11.15, sen-
sitivity was 50, specificity was 88, negative predic- Acknowledgement
tive value was 98 and accuracy rate was 88, re-
spectively (Table VII). Phelan et al. published
We are most grateful to Prof. Dr. H.P. van
that the false-negative rate was 1.9 and the posi-
Geijn for his advice and corrections to the paper
tive predictive value was 6.3% [23]. Brown and
in the preparatory phase.
Patrick concluded that negative predictive value
was 98.2 and positive predictive value was 85.7%
[241. Devoe et al. concluded that the negative References
predictive value was 92.8 and positive predictive
value was 30.6% [25]. Our negative predictive Rochard F, Schifrin BS, Goupil F et al. Nonstressed fetal
value is approximately the same in comparison heart monitoring in antepartum period. Am J Obstet Gy-
with the data available in literature but there are necol 1976;126:699.
Schifrin BS. Antepartum fetal heart rate monitoring. In:
some differences with regard to positive predic- Cluck L, ed. Intra-uterine asphyxia and developing fetal
tive value (11%). brain Chicago: Year Book Publishers Inc., 1977;205.
On the other hand, although OCT has a higher Kubli F, Ruttgers H, Hailer U et al. Die antepartale fetale
sensitivity rate (60%) and positive predictive value herzfrequenz II. Verhalteo-von Grundfrequenz. fluktua-
(18%) than NST (50 and ll%, respectively), there tion und dezelerationen bei antepartalem fruchttod. 2.
Geburtshilfe. Perinatol 1972;176:349.
is not any statistical difference in relation to Hammacher K, Huter KA, Bokelmann J, Werners PH.
positive predictive value between NST and OCT Fetal heart frequencies and condition of the fetus and
(P < 0.05). It means that there is no superiority newborn. Gynakologe 1968;166:349.
58

5 Hammacher K. The clinical significance of cardiotoco- 17 Manning FA, Lange IR, Morrison I, Harman CR. Fetal
graph In Perinatal Medicine. Huntingford PJ, Hunter ES, biophysical profile score and the non-stress test: a compar-
Saling E, eds. New York: Academic Press 1968;80. ative trial. Obstet Gynecol 1984;64:326.
6 Smith CU, Nguyen HN, Kovacs B, McCart D, Phelan JP, 18 Fleischer A, Schulman H, Farmakides G, Perrotta LA et
Paul RH. Fetal death following antepartum fetal heart al. Antepartum nonstress test and the postmature preg-
rate testing: a review 65 cases. Obstet Gynecol 1987;63:48,. nancy. Obstet Gynecol 1985;66:80.
7 Goldkrand JW, Benjamin DS. Antepartum fetal heart rate 19 Evertson LR, Gautheir RJ, Schifrin BS, Paul RH. An-
testing: a clinical appraisal. Obstet Gynecol 1985;66:88. tepartum fetal heart rate testing: evaluation of the non-
8 Freeman RK. The use of oxytocin challenge test for an- stress test. Am J Obstet Gynecol 1979;133:29.
tepartum clinic evaluation of uteroplacental respiratory 20 Manning FA, Platt LD, Sipos L, Keegan KA, Fetal breath-
function. Am J Obstet Gynecol 1975;124:481. ing movements and nonstress test in high risk pregnancy.
9 Schifrin B.S. The rational of antepartum fetal heart rate Am J Obstet Gynecol 1979;135:511.
monitoring. J Reprod Med 1979;23:213. 21 Schifrin BS. Antepartum fetal heart rate monitoring.
10 Martin CB, Schifrin BS. Prenatal fetal monitoring. In: Fourth International Symposium on Perinatal Medicine.
Aladjem S, Brown AK, eds. Perinatal intensive care. Co. Las Vegas. Nevada. 9-13 April 1979.
St. Louis: CV Mosby 1977;155. 22 Schifrin BS. Foye G. Amato J. Kates R. McKenna J.
11 Gabbe S, Freeman R, Goebelsman V. Evaluation of con- Routine fetal heart rate monitoring in the antepartum
traction stress test before 33 weeks gestation. Obstet Gy- period. Obstet Gynecol 1979;54:21.
necol 1978;52:649. 23 Phelan JP, Platt LD, Yeh S, Trujillo M, Paul RH. Contin-
12 Cetrolu CL, Nochimson NT. Freeman RK. Prolonged uing role of the nonstress test in the management of
pregnancy. Postgraduate course in clinic concepts in the postdates pregnancy. Obstet Gynecol 1984;64:624.
management of high-risk pregnancy. LAC-USC Medical 24 Brown R, Patrick J. The nonstress test: how long is enough
Center. Department of Ob/Gyn 1978. Am J Obstet Gynecol 1978;130:765.
13 Freeman RK, Anderson G, Dorchester W. A prospective 25 Devoe LD. Clinical implications of prospective antepar-
multinstitutional study of antepartum fetal heart rate mon- turn fetal heart rate testing. Am J Obstet Gynecol
itoring. 2. Contraction stress test versus non-stress for 1980;137:983.
primary surveillance. Am J Obstet Gynecol 1982;143:775. 26 Fox HE, Steinbrecher M, Ripton B. Antepartum fetal
14 Manning FA, Platt LD, Sipos L. Antepartum fetal evalua- heart rate and uterine activity studies. I. Preliminary re-
tion. Development of fetal biophysical profile. Am J Ob- port of acceleration and the oxytocin test. Am J Obstet
stet Gynecol 1980;136:787. Gynecol 1976;126:61.
15 Schifrin BS, Giintes (OcaklV, Gergely RC, Eden R, Roll 27 Thacker SB, Berkelman RL. Assessing the diagnostic ac-
K, Jacobs J. The role of real time scanning in antenatal curacy and efficacy of selected antepartum fetal surveil-
fetal surveillance. Am J Obstet Gynecol 1981;140:525. lance techniques. Obstet Gynecol Surv 1986;41-3:121.
16 Manning FA, Baskett TF, Morrison J et al. Fetal biophysi- 28 Huddleston JF, Sutcliff G, Camey FE et al. Oxytocin
cal profile scoring: a prospective study in 1184 high-risk challenge test for antepartum assessment. Am J Obstet
patient. Am J Obstet Gynecol 1981;140:189. Gynecol 1979;135:609.