Sterile water blocks for labor pain
Volume 164 Number 5, Part 1
tween labor pain and fetal position and the effect of analgesia on pelvic floor tone, cervical tension, contraction pattern, and fetal rotation are interesting and an evident subject for further studies. REFERENCES I. Melzack R, Schaffelberg D. Low-back pain during labor. AMJ OBSTET GYNECOL 1987;156:901-5. 2. Erkkola R, Pikkola P, Kanto J. Transcutaneous nerve stimulation for pain relief during labour: a controlled study. Ann Chir Gynaecol 1980;69:273-7. 3. Miller Jones CMH. Transcutaneous nerve stimulation in labour. Anaesthesia 1980;35:372-5. 4. Permin M, Ipsen L, Hansen MK, et al. Relief from pain localized to the lower back in early labour. Z Geburtshilfe PerinatoI1981;185:3601-3. 5. Ulmsten U, Sandahl B, Lundin C, Andersson KE. Treatment of labour pain with locally applied ketocaine. Acta Obstet Gynecol Scand 1980;59:209-12. 6. Rose D. Local anaesthesia in first and second stage labour. N EnglJ Med 1929;201:117-25.
7. Abrams AA. Obliteration of pain at the site of reference by intradermal infiltration anesthesia in first-stage labor. N Engl J Med 1950;243:636-40. 8. Bengtsson J, Worning AM, Gertz J, et al. Pain due to urolithiasis treated by intracutaneous injection of sterile water. Ugeskr Laeger 1981;143:3463-5. 9. Scott J, Huskisson EC. Graphic representation of pain. Pain 1976;2: 175-84. 10. Bonica JJ. Peripheral mechanisms and pathways of parturition pain. Br J Anaesth 1979;51:35-95. I I. Basbaum AI, Fields HL. Endogenous pain control mechanisms: review and hypothesis. Ann Neurol 1978;4:45161. 12. Sherman JE, Liebeskind JC. An endorphinergic, centrifugal substrate of pain modulation: recent findings, current concepts, and complexities. In: Bonica JJ, ed. Pain. New York: Raven Press, 1980:191-204. 13. Melzack R, Wall PD. Pain mechanism: a new theory. Science 1965;150:971-9. 14. Wall PD. The role of substantia gelatinosa as a gate control. In: Bonica JJ, ed. Pain. New York: Raven Press, 1980:205-31.
Electronic fetal heart monitoring, auscultation, and neonatal outcome Patricia H. Ellison, MD: Mark Foster, MA: Margaret Sheridan-Pereira, MD: and Dermot MacDonald, MDc Denver, Colorado, and Dublin, Ireland In a large randomized, controlled study of fetal heart rate monitoring with either continuous electronic fetal heart monitoring or auscultation at specified intervals, only one pattern of deviation in the fetal heart rate correlated significantly with neonatal neurologic examinations at 0 to 48 hours and 72 hours to 1 week: late decelerations in stage 1 and in stage 2. Other variables from labor and delivery, specifically, duration of labor after hospital admission, failure of labor to progress, number of fetal scalp pH values, and presence of meconium were important predictors of neonatal outcome in the regression analyses. The fetal heart rate deviations did contribute significantly to the percent variance accounted for in the regression analyses with neonatal outcomes of Apgar scores at 1 and 5 minutes and serial neonatal neurologic examinations. (AM J OBSTET GYNECOL 1991;164:1281-9.)
Key words: Continuous electronic fetal heart monitoring, late decelerations, Apgar scores at I and 5 minutes, neonatal neurologic examination Electronic fetal heart monitoring, when introduced in the 1960s and refined in subsequent years, brought hope that there could be a way to detect, in utero, an From the Department of Psychology, University of Denver," and Coombs Hospital' and National Maternity Hospital,' Dublin. Supported in part by the Presbyterian/ St. Luke's Foundation, Denver, Colorado. Received for publication March 6, 1990; revised December 3, 1990; accepted December 21, 1990. Reprint requests: Patricia H. Ellison, MD, Research Professor, Department of Psychology, University of Denver, Denver, CO 80208. 6/1 /27569
already compromised fetus, as well as a fetus in the process of deterioration, particularly during labor and delivery. In certain situations this technology has been fairly reliable, e.g., in the recognition of a dying fetus who previously might have been stillborn. It has been less reliable in the detection of fetuses with hypoxiaischemia of the central nervous system and secondarily decreased regulation of heart rate. In the majority of studies the relation was examined between fetal heart rate and the Apgar score, usually at I and 5 minutes. Very few studies have examined 1281
1282 Ellison at al.
May 1991
Am
J Obstet Gynecol
Table IA. Perinatal variables for four subsamples in percentage s A uscultati on plus Aus cultation (n = 2362)
Condition of mother Eclampsia Diabetes Failure of labor to progress T reatment Induction of labor Use of pethidine Use of epidural anesthesia Use of oxytocin Use of forceps Use of cesarean section Condition of fetus Postterm Breech Sampling of scalp pH Abnormal heart rate Presence of meconium Condition of neonate Stillborn Neonatal death Neo na tal convulsions
ElectroIi ic f etal heart monitoring (n = 23 62)
neurologic examiliation (n
= 135)
El ectronic f etal heart monitoring plus neurologic examina tion (n = 135)
1.6 1.0 5.8
1.9 1.2 9.0
6.6 3.6 11.6
6.8 6.8 15.5
16.8 46 .6 3.3 22.2 5.8 \.5
11.7 53 .8 3.8 28.3 13.8 2.4
26.8 46.4 3.6 23.9 11.6 7.2
18.9 54.7 2.7 41.9 25.0 1.4
5.5 \.8 0.9 I\.3
4.3 \.0 6.9 5.9 10.0
5.1 8.0 2.2 5.1 15.2
2.0 2.7 15.6 10.8 10.8
0.4 0.9 0.3
0.0 1.8 0.2
0.0 0.7 3.6
0.0 0.0 4.1
l.l
the relation between fetal heart monitoring and neonatal neurologic examination obtained in a standardized manner."? However, the gener al course of most compromised newborns is improvement from the minute of birth through the hours or days until such examinations are performed , especially with expert re suscitation and neonatal care. In this study we use the data from the Dublin randomized, controlled trial' to ex amine the relationships between a number of maternal and labor-and-delivery variables, including fetal heart rate and different fetal heart patterns, to several neonatal outcomes. The questions posed were: (1) Which spe cific fetal heart patterns are related to neonatal outcome as reflected in the Apgar scores at I and 5 minutes, the presence of convulsions, and the measures based on neurologic examination at 0 to 48 and 72 hours to I week ? (2) Which other labor-and-delivery variables are related to these neurologic outcomes? (3) How mu ch does each of these variables contribute to the prediction of outcome? (4) On which of these indicators do obstetricians base interventions, such as use of forceps or cesarean section? All newborns in this study were examined by a physician before discharge. Approximately 1000 were examined neurologically by a research physician using a standardized method. These newborns constituted all those admitted to the special care baby unit, all those with any neurologic abnormality on routine examination , and all those of special con cern to the n ursing staff. Every effort was made to examine in a stand ardized manner all newborns at risk for or suspected to have neurologic abnormality. The subsarnples available
for various analyses differed in size. However, all subsamples were not on ly larger than those in most such studies but sufficiently large to consider that the results were likely to represent the population well. Methods
The data in this study were those from the Dublin trial, a randomized, controlled study that compared the effectiveness of electronic fetal heart monitoring and auscultation in assessing the health of the fetus du r ing labor and delivery."T here were 12,964 women involved in the study at the National Maternity Hospital in Du blin between March 1981 and April 1983. Data for 10,089 deliveries were sent to the University of Denver for analyses. Women who had heavily meconiumstained or markedly decreased a mniotic fluid or who had fetuses with abnormal heart rates on admission were not enrolled. The women were randomly assigned to assessment by continuous electronic fetal heart monitoring or to intermittent monitoring by auscultation . Fetal scalp sampling was available to both sections. Four subsamples were used in the re sults described in thi s article. The auscultation-only group was a random sa m ple of all deliveries for which data were sent and in which only auscultation of the fetal heart was performed (n = 2362). The electronic fetal heart monitoring group, which was equal in number to the au scultation-only group, included all deliveries assigned to electronic fetal heart monitoring for which dat a were avai lable for both stages I and 2 (n = 2362). The auscultation-plus-neurologic group, a random sample equal to that of those babies who ha d electronic fetal heart monitoring in stages I and 2 and a
Fetal heart monitoring and neonatal outcomes
Volume 164 Number 5, Part 1
1283
Table lB. Mean Apgar scores and duration of labor
I min 5 min Duration of labor (after entering hospital)
Auscultation (n = 2362)
Electronic fetal heart monitoring (n = 2362)
Auscultation plus neurulugic examination (n = 135)
8.8 ± 1.0 9.1 ± 0.6 3.5 ± 2.6
8.7 ± 1.0 9.1 ± 0.6 4.2 ± 2.6
7.6 ± 2.4 8.7 ± l.l 3.7 ± 2.8
neurologic evaluation at 0 to 48 hours and 72 hours to 1 week, was from those deliveries that were monitored only by auscultation and in which the babies underwent a neurologic examination by the research physician at both 0 to 48 hours and 72 hours to 1 week (n = 135). The electronic fetal heart monitoring-plusneurologic group, equal in number to the auscultationplus-neurologic group, included those neonates whose mothers were assigned to electronic fetal heart monitoring, had such data during stages 1 and 2, and had both neurologic examinations by the research physician (n = 135). We wanted to compare samples of equal size and used random sampling to form the subsamples. In the regression analyses the sample sizes of each of these groups decreased slightly because of missing data. A large number of maternal, fetal, and labor-anddelivery variables were recorded, including maternal age, parity, gravidity, gestational age of fetus, use of oxytocin or pethidine, length of time from admission to delivery, failure of labor to progress, delivery by cesarean section or forceps, and Apgar scores at 1 and 5 minutes. The fetal heart tracings were interpreted by an obstetrician who had no knowledge about the mother, the other information about the labor and delivery, or the condition of the baby. The categories were marked bradycardia, late deceleration pattern, severe variable deceleration pattern, marked tachycardia, moderate tachycardia with reduced variability, moderate bradycardia with reduced variability, minimal variability, moderate variable decelerations, moderate tachycardia with normal variability, mild variable deceleration pattern, early deceleration pattern, reduced variability, moderate bradycardia with normal variability, acceleration pattern, normal baseline and variability, and unclassifiable (poor quality, labor too short, monitor disconnected, and not known-trace missing). All newborns were examined physically and neurologically before discharge by a physician. Those newborns for whom there was concern about neurologic dysfunction by physician or nurse for any reason or who were admitted to the neonatal intensive care unit were examined by one physician using a comprehensive neonatal neurologic examination. This included individually scored items from the Dubowitz," Prechtl, II
Electronic fetal heart monitoring plus neurologic examination (n = 135)
7.6 ± 2.1 8.7 ± 1.2 4.9 ± 3.1
and Brazelton" methods of evaluation, as well as items for gestational age scoring. The babies were examined at 0 to 48 hours, at 72 hours to 1 week, and at weekly intervals until discharge. The overall examination result was scored as normal, questionably abnormal, or abnormal. Data also were recorded about neonatal death, congenital abnormalities, and neonatal convulsions. The neurologic data were subsequently analyzed extensively and used to construct a neonatal neurologic assessment instrument with a total score and seven subscores. I" Frequencies were reviewed for all variables, as well as distributions and skews. Pearson product moment correlations and point biserial correlations (for dichotomous variables) were obtained and reviewed among all variables for each sample. Multiple regression analyses were obtained for different outcomes: Apgar scores at 1 and 5 minutes and neurologic examination at 0 to 48 hours and 72 hours to 1 week. The variables entered in the regression analyses for the two auscultation groups were confined to duration of labor, presence or absence of meconium, failure of labor to progress, use of oxytocin, number of fetal scalp samples, and any fetal heart rate abnormality. The regression analyses for the two electronic fetal heart monitoring groups were confined to the first five variables as noted above plus the ominous (marked bradycardia, late deceleration pattern, and severe variable deceleration pattern) and suspicious (marked tachycardia, moderate tachycardia with reduced variability, moderate bradycardia with reduced variability, minimal variability, and moderate variable decelerations) fetal heart patterns entered separately for stages I and 2. The ominous and suspicious rankings were determined by the obstetric classifications. The decision to select variables for entry in the regression analyses were based on review of the correlation matrices. Finally, regression analyses were obtained to determine those labor variables that accounted for obstetric intervention-use of forceps or cesarean section. The variables entered in the regression analyses were restricted to those described above for the four samples. Results
The frequencies of key variables including eclampsia, diabetes, and breech position for the several subsam-
1284 Ellison el al. Am
May 1991 Gyneco l
J O bstet
Table II. Frequencies of specific fetal heart monitoring patt erns in pe rcent Stage I
E lectronic fetal heart monitoring
Marked brad ycardia Late decelerations Severe variable deceleration s Marked tach ycardia Moderate tach ycardia with reduced variability Mode rate brad ycardia with reduced variability Minimal variability Moderate variable decelerations Moderat e tach ycardia with normal vari abilit y Mild var iable decelerati on s Earl y de cele ration pattern Reduced variability Mod erate br ad ycardia with normal var iability Acceleration pattern No r mal baseline and vari abilit y Unclassifiable
Stage 2 Electronic f etal heart monitoring plus neurologic examination
El ectronic f etal heart monitoring
E lectronic fetal heart monitoring plus neurologic examination
14.0 4.7 2.0 0 0
12.8 10.1 2.0 0 0
1.5 1.8 1.4 0.4 0.2
0.7 3.4 3.4 1.4 2.0
O
0
0
0
0.4 0.9 3.2
0.7 0.7 5.4
0 2.1 1.1
0 2.7 2.7
15.3 11.2 3.0 10.7
20.3 15.5 3.4 6.1
12.6 39 .5 0.4 6.6
14.9 33 . 1 0.7 5.4
4.5 45.6 0
4.1 33.1 0
0.5 16.4 0
0.7 13.5 1.4
pies are shown in Table lA o Intervention by use of forceps is low, with the ex ception of the ele ctronic fetal heart monitoring-plus-neurologic examination subsample. The rate of cesarean section is also low, particularly in comparison with that of American samples . The means and standard d eviations for Apgar scores at 1 and 5 minutes and for the duration of labor after hospital en try are shown in Table lB . The 1- and 5minute Apgar scores were somewhat lower and the standard deviations larger for the two subs amples with neuro logic examination . T he duration of labor was longest for the electronic fet al heart monitoring-plusneurologic examination subsample. The frequencies of fetal he art monitoring patterns are shown in Table II for the sub set with electronic monitoring. The frequencies for the patterns th at hav e been characterized as ominous are low, with the exception of marked bradycardia in stage 2 lab or. The frequency of patterns that have been cha racterized as "non reassu ring" is high during stage 2 labor, especially for earl y deceleration. The freque ncy of normal baseline and vari abilit y decreases to only a sm all percentage of fetuses by stage 2. Thus the majority of patterns were de scribed as "other th an normal." In Table lIlA the few correlations with significance are shown between those fetal heart patterns and outcomes. The most significant correlations were those between late de celerations (stage 1) and th e de velopment of convulsions and two neurologic examinations. There also were statistically significant correlations between late d ecelerations (stage 2) and the develop men t of convulsions and two neonatal neurologic evaluations, but they wer e smaller. The cor relations bet ween presence
of meconium, duration of labor, failure to progress, use of oxytocin, and number of scalp pH values and the several ou tcomes are sho wn in Table IIIB. The onl y sign ifican t correlation with the second neurologic evaluation is presence of me conium. In Table IV the results are shown for the regression analyses with Apgar score at 1 minute as the dependent variable. In all the subsamples the percent variance accounted for ranged from 3.2 to 24.3. The heart rate abnormalities were significant in adding to the predictio n in the auscultation, electronic fet al heart mon itoring, and electronic fet al heart monitoring-plusneurologic examination gro u ps. However, larger contributions wer e made when specific patterns were used as in the ele ctronic fet al he art monitoring and ele ctronic fet al heart monitoring-plus-neurologic examination groups. The specific patterns combined to make the largest contribution in the electronic fetal heart monitoring-plus-neurologic examination grou p; those patterns were minimal variability (stage I), mar ked bradycardia (stage 2), and late dec elerations (stage 2). In the electronic fetal he art monitoring group, late deceleration s (stage 2) was the most powerful predictor var iable ; in the electronic fetal heart monitoring-plus-neurologic examination group, marked bradycardia (stage 2) was th e lead or most powerful predictive vari ab le. In Table V the results a re shown for the regression analyses with Apgar score at 5 minutes as the dependent variable . In the aus cultation and ele ctronic fetal hear t monitoring groups extremely sma ll amou n ts of the variance were accounted for. In the subsamples evaluated neurologically, a larger percent of the vari-
Fetal heart monitoring and neonatal outcomes
Volume 164
1285
Number 5, Part I
Table IlIA. Correlations of specific fetal heart patterns to outcomes Electronic fetal heart monitoring plus neurologic examination
Electronic fetal heart monitoring
Neurologic examinations
Stage I Late decelerations Severe variable decelerations Marked tachycardia Moderate variable decelerations Early decelerations Normal baseline and variability Stage 2 Late decelerations Early decelerations
Apgar score at 1 min (n = 2362)
Apgar score at 5 min (n = 2362)
Convulsions (n = 135)
-0.07* - 0.11 *
-0.04t -0.07*
0.38* -0.04
-0.04t -0.04t
-0.05t 0.01
-0.02 -0.02
-0.04 -0.04
-0.02 -0.02
-0.05t 0.06:j:
0.02 0.04t
0.01 -0.05
-0.09 -0.15
-0.07 -0.05
-0.14* 0.04t
-0.06:j: 0.02
0.32* -0.14t
0.27:j: -0.09
0.26:j: -0.15
Only those correlations are shown for which at least one correlation had a *P < 0.001. tp < 0.05. :j:p < 0.01.
p
Volume 164 Number 5, Part 1
tween labor pain and fetal position and the effect of analgesia on pelvic floor tone, cervical tension, contraction pattern, and fetal rotation are interesting and an evident subject for further studies. REFERENCES I. Melzack R, Schaffelberg D. Low-back pain during labor. AMJ OBSTET GYNECOL 1987;156:901-5. 2. Erkkola R, Pikkola P, Kanto J. Transcutaneous nerve stimulation for pain relief during labour: a controlled study. Ann Chir Gynaecol 1980;69:273-7. 3. Miller Jones CMH. Transcutaneous nerve stimulation in labour. Anaesthesia 1980;35:372-5. 4. Permin M, Ipsen L, Hansen MK, et al. Relief from pain localized to the lower back in early labour. Z Geburtshilfe PerinatoI1981;185:3601-3. 5. Ulmsten U, Sandahl B, Lundin C, Andersson KE. Treatment of labour pain with locally applied ketocaine. Acta Obstet Gynecol Scand 1980;59:209-12. 6. Rose D. Local anaesthesia in first and second stage labour. N EnglJ Med 1929;201:117-25.
7. Abrams AA. Obliteration of pain at the site of reference by intradermal infiltration anesthesia in first-stage labor. N Engl J Med 1950;243:636-40. 8. Bengtsson J, Worning AM, Gertz J, et al. Pain due to urolithiasis treated by intracutaneous injection of sterile water. Ugeskr Laeger 1981;143:3463-5. 9. Scott J, Huskisson EC. Graphic representation of pain. Pain 1976;2: 175-84. 10. Bonica JJ. Peripheral mechanisms and pathways of parturition pain. Br J Anaesth 1979;51:35-95. I I. Basbaum AI, Fields HL. Endogenous pain control mechanisms: review and hypothesis. Ann Neurol 1978;4:45161. 12. Sherman JE, Liebeskind JC. An endorphinergic, centrifugal substrate of pain modulation: recent findings, current concepts, and complexities. In: Bonica JJ, ed. Pain. New York: Raven Press, 1980:191-204. 13. Melzack R, Wall PD. Pain mechanism: a new theory. Science 1965;150:971-9. 14. Wall PD. The role of substantia gelatinosa as a gate control. In: Bonica JJ, ed. Pain. New York: Raven Press, 1980:205-31.
Electronic fetal heart monitoring, auscultation, and neonatal outcome Patricia H. Ellison, MD: Mark Foster, MA: Margaret Sheridan-Pereira, MD: and Dermot MacDonald, MDc Denver, Colorado, and Dublin, Ireland In a large randomized, controlled study of fetal heart rate monitoring with either continuous electronic fetal heart monitoring or auscultation at specified intervals, only one pattern of deviation in the fetal heart rate correlated significantly with neonatal neurologic examinations at 0 to 48 hours and 72 hours to 1 week: late decelerations in stage 1 and in stage 2. Other variables from labor and delivery, specifically, duration of labor after hospital admission, failure of labor to progress, number of fetal scalp pH values, and presence of meconium were important predictors of neonatal outcome in the regression analyses. The fetal heart rate deviations did contribute significantly to the percent variance accounted for in the regression analyses with neonatal outcomes of Apgar scores at 1 and 5 minutes and serial neonatal neurologic examinations. (AM J OBSTET GYNECOL 1991;164:1281-9.)
Key words: Continuous electronic fetal heart monitoring, late decelerations, Apgar scores at I and 5 minutes, neonatal neurologic examination Electronic fetal heart monitoring, when introduced in the 1960s and refined in subsequent years, brought hope that there could be a way to detect, in utero, an From the Department of Psychology, University of Denver," and Coombs Hospital' and National Maternity Hospital,' Dublin. Supported in part by the Presbyterian/ St. Luke's Foundation, Denver, Colorado. Received for publication March 6, 1990; revised December 3, 1990; accepted December 21, 1990. Reprint requests: Patricia H. Ellison, MD, Research Professor, Department of Psychology, University of Denver, Denver, CO 80208. 6/1 /27569
already compromised fetus, as well as a fetus in the process of deterioration, particularly during labor and delivery. In certain situations this technology has been fairly reliable, e.g., in the recognition of a dying fetus who previously might have been stillborn. It has been less reliable in the detection of fetuses with hypoxiaischemia of the central nervous system and secondarily decreased regulation of heart rate. In the majority of studies the relation was examined between fetal heart rate and the Apgar score, usually at I and 5 minutes. Very few studies have examined 1281
1282 Ellison at al.
May 1991
Am
J Obstet Gynecol
Table IA. Perinatal variables for four subsamples in percentage s A uscultati on plus Aus cultation (n = 2362)
Condition of mother Eclampsia Diabetes Failure of labor to progress T reatment Induction of labor Use of pethidine Use of epidural anesthesia Use of oxytocin Use of forceps Use of cesarean section Condition of fetus Postterm Breech Sampling of scalp pH Abnormal heart rate Presence of meconium Condition of neonate Stillborn Neonatal death Neo na tal convulsions
ElectroIi ic f etal heart monitoring (n = 23 62)
neurologic examiliation (n
= 135)
El ectronic f etal heart monitoring plus neurologic examina tion (n = 135)
1.6 1.0 5.8
1.9 1.2 9.0
6.6 3.6 11.6
6.8 6.8 15.5
16.8 46 .6 3.3 22.2 5.8 \.5
11.7 53 .8 3.8 28.3 13.8 2.4
26.8 46.4 3.6 23.9 11.6 7.2
18.9 54.7 2.7 41.9 25.0 1.4
5.5 \.8 0.9 I\.3
4.3 \.0 6.9 5.9 10.0
5.1 8.0 2.2 5.1 15.2
2.0 2.7 15.6 10.8 10.8
0.4 0.9 0.3
0.0 1.8 0.2
0.0 0.7 3.6
0.0 0.0 4.1
l.l
the relation between fetal heart monitoring and neonatal neurologic examination obtained in a standardized manner."? However, the gener al course of most compromised newborns is improvement from the minute of birth through the hours or days until such examinations are performed , especially with expert re suscitation and neonatal care. In this study we use the data from the Dublin randomized, controlled trial' to ex amine the relationships between a number of maternal and labor-and-delivery variables, including fetal heart rate and different fetal heart patterns, to several neonatal outcomes. The questions posed were: (1) Which spe cific fetal heart patterns are related to neonatal outcome as reflected in the Apgar scores at I and 5 minutes, the presence of convulsions, and the measures based on neurologic examination at 0 to 48 and 72 hours to I week ? (2) Which other labor-and-delivery variables are related to these neurologic outcomes? (3) How mu ch does each of these variables contribute to the prediction of outcome? (4) On which of these indicators do obstetricians base interventions, such as use of forceps or cesarean section? All newborns in this study were examined by a physician before discharge. Approximately 1000 were examined neurologically by a research physician using a standardized method. These newborns constituted all those admitted to the special care baby unit, all those with any neurologic abnormality on routine examination , and all those of special con cern to the n ursing staff. Every effort was made to examine in a stand ardized manner all newborns at risk for or suspected to have neurologic abnormality. The subsarnples available
for various analyses differed in size. However, all subsamples were not on ly larger than those in most such studies but sufficiently large to consider that the results were likely to represent the population well. Methods
The data in this study were those from the Dublin trial, a randomized, controlled study that compared the effectiveness of electronic fetal heart monitoring and auscultation in assessing the health of the fetus du r ing labor and delivery."T here were 12,964 women involved in the study at the National Maternity Hospital in Du blin between March 1981 and April 1983. Data for 10,089 deliveries were sent to the University of Denver for analyses. Women who had heavily meconiumstained or markedly decreased a mniotic fluid or who had fetuses with abnormal heart rates on admission were not enrolled. The women were randomly assigned to assessment by continuous electronic fetal heart monitoring or to intermittent monitoring by auscultation . Fetal scalp sampling was available to both sections. Four subsamples were used in the re sults described in thi s article. The auscultation-only group was a random sa m ple of all deliveries for which data were sent and in which only auscultation of the fetal heart was performed (n = 2362). The electronic fetal heart monitoring group, which was equal in number to the au scultation-only group, included all deliveries assigned to electronic fetal heart monitoring for which dat a were avai lable for both stages I and 2 (n = 2362). The auscultation-plus-neurologic group, a random sample equal to that of those babies who ha d electronic fetal heart monitoring in stages I and 2 and a
Fetal heart monitoring and neonatal outcomes
Volume 164 Number 5, Part 1
1283
Table lB. Mean Apgar scores and duration of labor
I min 5 min Duration of labor (after entering hospital)
Auscultation (n = 2362)
Electronic fetal heart monitoring (n = 2362)
Auscultation plus neurulugic examination (n = 135)
8.8 ± 1.0 9.1 ± 0.6 3.5 ± 2.6
8.7 ± 1.0 9.1 ± 0.6 4.2 ± 2.6
7.6 ± 2.4 8.7 ± l.l 3.7 ± 2.8
neurologic evaluation at 0 to 48 hours and 72 hours to 1 week, was from those deliveries that were monitored only by auscultation and in which the babies underwent a neurologic examination by the research physician at both 0 to 48 hours and 72 hours to 1 week (n = 135). The electronic fetal heart monitoring-plusneurologic group, equal in number to the auscultationplus-neurologic group, included those neonates whose mothers were assigned to electronic fetal heart monitoring, had such data during stages 1 and 2, and had both neurologic examinations by the research physician (n = 135). We wanted to compare samples of equal size and used random sampling to form the subsamples. In the regression analyses the sample sizes of each of these groups decreased slightly because of missing data. A large number of maternal, fetal, and labor-anddelivery variables were recorded, including maternal age, parity, gravidity, gestational age of fetus, use of oxytocin or pethidine, length of time from admission to delivery, failure of labor to progress, delivery by cesarean section or forceps, and Apgar scores at 1 and 5 minutes. The fetal heart tracings were interpreted by an obstetrician who had no knowledge about the mother, the other information about the labor and delivery, or the condition of the baby. The categories were marked bradycardia, late deceleration pattern, severe variable deceleration pattern, marked tachycardia, moderate tachycardia with reduced variability, moderate bradycardia with reduced variability, minimal variability, moderate variable decelerations, moderate tachycardia with normal variability, mild variable deceleration pattern, early deceleration pattern, reduced variability, moderate bradycardia with normal variability, acceleration pattern, normal baseline and variability, and unclassifiable (poor quality, labor too short, monitor disconnected, and not known-trace missing). All newborns were examined physically and neurologically before discharge by a physician. Those newborns for whom there was concern about neurologic dysfunction by physician or nurse for any reason or who were admitted to the neonatal intensive care unit were examined by one physician using a comprehensive neonatal neurologic examination. This included individually scored items from the Dubowitz," Prechtl, II
Electronic fetal heart monitoring plus neurologic examination (n = 135)
7.6 ± 2.1 8.7 ± 1.2 4.9 ± 3.1
and Brazelton" methods of evaluation, as well as items for gestational age scoring. The babies were examined at 0 to 48 hours, at 72 hours to 1 week, and at weekly intervals until discharge. The overall examination result was scored as normal, questionably abnormal, or abnormal. Data also were recorded about neonatal death, congenital abnormalities, and neonatal convulsions. The neurologic data were subsequently analyzed extensively and used to construct a neonatal neurologic assessment instrument with a total score and seven subscores. I" Frequencies were reviewed for all variables, as well as distributions and skews. Pearson product moment correlations and point biserial correlations (for dichotomous variables) were obtained and reviewed among all variables for each sample. Multiple regression analyses were obtained for different outcomes: Apgar scores at 1 and 5 minutes and neurologic examination at 0 to 48 hours and 72 hours to 1 week. The variables entered in the regression analyses for the two auscultation groups were confined to duration of labor, presence or absence of meconium, failure of labor to progress, use of oxytocin, number of fetal scalp samples, and any fetal heart rate abnormality. The regression analyses for the two electronic fetal heart monitoring groups were confined to the first five variables as noted above plus the ominous (marked bradycardia, late deceleration pattern, and severe variable deceleration pattern) and suspicious (marked tachycardia, moderate tachycardia with reduced variability, moderate bradycardia with reduced variability, minimal variability, and moderate variable decelerations) fetal heart patterns entered separately for stages I and 2. The ominous and suspicious rankings were determined by the obstetric classifications. The decision to select variables for entry in the regression analyses were based on review of the correlation matrices. Finally, regression analyses were obtained to determine those labor variables that accounted for obstetric intervention-use of forceps or cesarean section. The variables entered in the regression analyses were restricted to those described above for the four samples. Results
The frequencies of key variables including eclampsia, diabetes, and breech position for the several subsam-
1284 Ellison el al. Am
May 1991 Gyneco l
J O bstet
Table II. Frequencies of specific fetal heart monitoring patt erns in pe rcent Stage I
E lectronic fetal heart monitoring
Marked brad ycardia Late decelerations Severe variable deceleration s Marked tach ycardia Moderate tach ycardia with reduced variability Mode rate brad ycardia with reduced variability Minimal variability Moderate variable decelerations Moderat e tach ycardia with normal vari abilit y Mild var iable decelerati on s Earl y de cele ration pattern Reduced variability Mod erate br ad ycardia with normal var iability Acceleration pattern No r mal baseline and vari abilit y Unclassifiable
Stage 2 Electronic f etal heart monitoring plus neurologic examination
El ectronic f etal heart monitoring
E lectronic fetal heart monitoring plus neurologic examination
14.0 4.7 2.0 0 0
12.8 10.1 2.0 0 0
1.5 1.8 1.4 0.4 0.2
0.7 3.4 3.4 1.4 2.0
O
0
0
0
0.4 0.9 3.2
0.7 0.7 5.4
0 2.1 1.1
0 2.7 2.7
15.3 11.2 3.0 10.7
20.3 15.5 3.4 6.1
12.6 39 .5 0.4 6.6
14.9 33 . 1 0.7 5.4
4.5 45.6 0
4.1 33.1 0
0.5 16.4 0
0.7 13.5 1.4
pies are shown in Table lA o Intervention by use of forceps is low, with the ex ception of the ele ctronic fetal heart monitoring-plus-neurologic examination subsample. The rate of cesarean section is also low, particularly in comparison with that of American samples . The means and standard d eviations for Apgar scores at 1 and 5 minutes and for the duration of labor after hospital en try are shown in Table lB . The 1- and 5minute Apgar scores were somewhat lower and the standard deviations larger for the two subs amples with neuro logic examination . T he duration of labor was longest for the electronic fet al heart monitoring-plusneurologic examination subsample. The frequencies of fetal he art monitoring patterns are shown in Table II for the sub set with electronic monitoring. The frequencies for the patterns th at hav e been characterized as ominous are low, with the exception of marked bradycardia in stage 2 lab or. The frequency of patterns that have been cha racterized as "non reassu ring" is high during stage 2 labor, especially for earl y deceleration. The freque ncy of normal baseline and vari abilit y decreases to only a sm all percentage of fetuses by stage 2. Thus the majority of patterns were de scribed as "other th an normal." In Table lIlA the few correlations with significance are shown between those fetal heart patterns and outcomes. The most significant correlations were those between late de celerations (stage 1) and th e de velopment of convulsions and two neurologic examinations. There also were statistically significant correlations between late d ecelerations (stage 2) and the develop men t of convulsions and two neonatal neurologic evaluations, but they wer e smaller. The cor relations bet ween presence
of meconium, duration of labor, failure to progress, use of oxytocin, and number of scalp pH values and the several ou tcomes are sho wn in Table IIIB. The onl y sign ifican t correlation with the second neurologic evaluation is presence of me conium. In Table IV the results are shown for the regression analyses with Apgar score at 1 minute as the dependent variable. In all the subsamples the percent variance accounted for ranged from 3.2 to 24.3. The heart rate abnormalities were significant in adding to the predictio n in the auscultation, electronic fet al heart mon itoring, and electronic fet al heart monitoring-plusneurologic examination gro u ps. However, larger contributions wer e made when specific patterns were used as in the ele ctronic fet al he art monitoring and ele ctronic fet al heart monitoring-plus-neurologic examination groups. The specific patterns combined to make the largest contribution in the electronic fetal heart monitoring-plus-neurologic examination grou p; those patterns were minimal variability (stage I), mar ked bradycardia (stage 2), and late dec elerations (stage 2). In the electronic fetal he art monitoring group, late deceleration s (stage 2) was the most powerful predictor var iable ; in the electronic fetal heart monitoring-plus-neurologic examination group, marked bradycardia (stage 2) was th e lead or most powerful predictive vari ab le. In Table V the results a re shown for the regression analyses with Apgar score at 5 minutes as the dependent variable . In the aus cultation and ele ctronic fetal hear t monitoring groups extremely sma ll amou n ts of the variance were accounted for. In the subsamples evaluated neurologically, a larger percent of the vari-
Fetal heart monitoring and neonatal outcomes
Volume 164
1285
Number 5, Part I
Table IlIA. Correlations of specific fetal heart patterns to outcomes Electronic fetal heart monitoring plus neurologic examination
Electronic fetal heart monitoring
Neurologic examinations
Stage I Late decelerations Severe variable decelerations Marked tachycardia Moderate variable decelerations Early decelerations Normal baseline and variability Stage 2 Late decelerations Early decelerations
Apgar score at 1 min (n = 2362)
Apgar score at 5 min (n = 2362)
Convulsions (n = 135)
-0.07* - 0.11 *
-0.04t -0.07*
0.38* -0.04
-0.04t -0.04t
-0.05t 0.01
-0.02 -0.02
-0.04 -0.04
-0.02 -0.02
-0.05t 0.06:j:
0.02 0.04t
0.01 -0.05
-0.09 -0.15
-0.07 -0.05
-0.14* 0.04t
-0.06:j: 0.02
0.32* -0.14t
0.27:j: -0.09
0.26:j: -0.15
Only those correlations are shown for which at least one correlation had a *P < 0.001. tp < 0.05. :j:p < 0.01.
p
