- Email: [email protected]
Antepartum surveillance for a history of stillbirth: When to begin?
February 1995 Am J Obstet Gynecol
Weeks eta[.
20. Fine RN. Diagnosis and treatment of fetal urinary tract abnormalities. J Pediatr 1992; 121:333-41. 21. Roberts JA, Kaack MB, Morvant AB. Vesicoureteral reflux in the primate. IV. Infection as cause of prolonged highgrade reflux. Pediatrics 1988;82:91-5. 22. Najmaldin A, Burge DM, Atwell JD. Reflux nephropathy secondary to intrauterine vesicoureteric reflux. J Pediatr Surg 1990;25:387-90. 23. Steele BT, Robitaille P, De Maria J, Grignon A. Follow-up
evaluation of prenatally recognised vesicoureteric reflux. J Pediatr 1989;115;95-6. 24. Gordon AC, Thomas DFM, Arthur RJ, Irving HC, Smith SEW. Prenatally diagnosed reflux: a follow-up study. Br J Urol 1990;65:407-12. 25. Zerin JM, Ritchey ML, Chang ACH. Incidental vesicoureteral reflux in neonates with antenatally detected hydronephrosis and other renal abnormalities. Radiology 1993;187:157-60.
Antepartum surveillance for a history of stillbirth: When to begin? Jonathan W. Weeks, MD," Tamerou Asrat, MD, b Mark A. Morgan, MD, b Michael Nageotte, MD, b Steven J. Thomas, MD, c and Roger K. Freeman, MD b Louisville, Kentucky, and Irvine, California OBJECTIVE: A history of stillbirth is universally accepted as an indication for antepartum fetal heart rate testing. Our goal was to examine when fetal testing should begin in an otherwise healthy patient with a history of stillbirth. STUDY DESIGN: This is a nonconcurrent cohort study of patients who were seen for antepartum surveillance from January 1979 to December 1991 with a history of stillbirth as the only indication for testing. Subsequent pregnancies were evaluated for adverse outcomes and abnormal antepartum test results. RESULTS: There was one case of recurrent stillbirth among the 300 study patients. Nineteen patients (6.4%) had one or more positive antepartum surveillance tests (positive contraction stress test or biophysical profile -<4). Three patients (1%) had positive tests before 32 weeks, all of whom were subsequently delivered without incident at term. Three patients were delivered for positive tests at < 36 weeks, one by cesarean section for fetal distress. We could not detect a relationship between the gestational age of the previous stillborn and the incidence of abnormal tests or fetal distress in subsequent pregnancies. CONCLUSION: Antepartum surveillance should begin at ->32 weeks in the healthy pregnant woman with a history of stillbirth. (AM J OBSTETGYNECOL1995;172:486-92.)
Key words: Stillbirth, antepartum surveillance, fetal heart rate testing, fetal monitoring
A history of stillbirth has long been recognized as a significant risk factor for subsequent poor pregnancy outcome. The perinatal collaborative studies in the United States and Britain demonstrated an increased risk of perinatal mortality and neurologic damage in subsequent births?' 2 Since the completion of these landmark studies there has been only one United States From the Department of Obstetrics and Gynecology, University of Louisville, a and the Department of Obstetrics and Gynecologyband the Division of Perinatology/ University of California, Irvine. Received for publication January 26, 1994; revised April 12, 1994; acceptedJuly 21, 1994. Reprint requests:Jonathan W. Weeks, MD, Department of Obstetrics and Gynecology, 2nd floor Ambulatory Care Bldg, 550 S. Jackson St., Louisville, KY 40292. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/59397 486
study to evaluate the significance of a history of stillbirth in the era of electronic fetal surveillance. Freeman et al? demonstrated that a history of stillbirth remains a significant risk factor and concluded that such patients deserve antepartum fetal testing. However, the p r o p e r time to initiate fetal monitoring remains undetermined. Although some sources r e c o m m e n d beginning at 32 to 34 weeks, 4 many clinicians initiate monitoring 2 to 4 weeks before the gestational age of the previous stillbirth. In part, this controversy stems from the paucity of published data on the timing of fetal monitoring in patients with a previous stillbirth. Approximately half of all stillbirths are the result of significant maternal medical conditions, the most common being hypertensive disorders and insulin-dependent diabetes. When to initiate antepartum surveillance
Volume 172, Number 2, Part 1 AmJ Obstet Gynecol
in such patients has been studied and r e p o r t e d in two prior publications from this institutionJ' 6 We now report our experience with antepartum surveillance in the otherwise healthy patient with a previous stillbirth. These patients were evaluated for adverse pregnancy outcome and antepartum tests results indicative of uteroplacental insufficiency. Our goal was to determine when to initiate antepartum fetal surveillance in healthy patients with a history of stillbirth. Material and methods All study patients were seen in the Centers for Fetal Evaluation at Long Beach Memorial and the University of California, Irvine, medical centers. The Long Beach Memorial antepartum testing computer database was reviewed from January 1979 through December 1991. A similar database at the University Of California, Irvine, was analyzed for patients tested from January 1988 through December 1991. During this period approximately 70,000 tests were performed on > 15,000 patients. The databases were queried for patients who were referred with singleton pregnancies and a history ,, of stillbirth as the only indication for testing. Medical records were reviewed to confirm the accuracy of the coded referral diagnosis. Patients with a history of intrauterine fetal death at < 20 weeks' gestation were excluded. The search yielded 39 patients from the University of California, Irvine, and 261 from Long Beach Memorial Medical Center. Approximately 15 patients from the Long Beach Memorial database had been r e p o r t e d in a previous collaborative study on antepartum fetal heart rate (FHR) testing? Patient histories, antepartum fetal test results, and pregnancy outcomes were prospectively collected and retrospectively analyzed. Recorded patient history included maternal age, gravidity, parity, gestational age when tested, testing method, test results, gestational age at delivery, indication for delivery, other medical or obstetric diagnoses, neonatal Apgar scores, neonatal intensive care admission, and birth weight. Neonates with birth weights < 10th percentile by California growth chart standards were considered to be growth restricted. 7 Obstetric or medical complications associated with the previous stillbirth were ascertained by reviewing the medical records. We placed an emphasis on describing the clinical setting in which the previous stillbirths occurred rather than relying on the death certificate and autopsy-derived classification systems that have been previously described. 8' 9 Autopsy reports were used in cases where the stillborn infants were delivered at the study institutions. When the cause of the previous stillbirth was stated to be an event that resulted in umbilical cord occlusion (e.g., prolapse, torsion, thrombosis, multiple tight loops, true knots), it was classified as a "cord accident." Pregnancy-induced hypertension was the assigned cause when an otherwise unexplained,
Weeks et al. 487
Table I. Causes of previous stillbirth Unknown Cord accident Infection Anomalies Abruptio placentae PIH Postdates Twins Trauma Intrapartum Drugs or alcohol IUGR Placental infarct Uterine rupture Other
147 48 13 13 14 12 10 9 6 6 5 5 4 1 7
(49.0) (16.0) (4.3) (4.3) (4.6) (4.0) (3.3) (3.0) (2.0) (2.0) (1.6) (1.6) (1.3) (0.3) (2.3)
PIH, Pregnancy-induced hypertension; IUGR, intrauterine growth retardation.
structurally normal stillborn infant occurred in a mother with signs and symptoms of pregnancy-induced hypertension or preeclampsia. Infection was considered to be the cause when there was funisitis or pneumonia on histopathologic study documented perinatal infection (e.g., syphilis or listeria), or when the mother was admitted with classic signs and symptoms of chorioamnionitis. Up until December 1989 the primary method of surveillance was weekly contraction stress tests (CST). Testing procedures and test interpretation have been described elsewhere. 4 From January 1990 through December 1991 the primary mode of surveillance was changed to a semiweekly modified biophysical profile (nonstress test [NST] and amniotic fluid index)J °' ~1 Tests were read as reactive if there were two accelerations that reached a peak of 15 beats/min above the baseline and lasted 15 seconds in a 20-minute window of recording time. The amniotic fluid index was performed as described by Phelan et alJ 2 If the NST was nonreactive or had FHR deceleration or if the amniotic fluid index was <5.0 cm, a CST or full biophysical profile was performed as "backup" for the abnormal primary test. Backup tests were performed in all instances of abnormal primary tests even if immediate delivery was planned. Intervention because of an abnormal test result was the decision of the primary physician, often in consultation with the perinatal staff. All FHR tests were prospectively read by at least one maternal-fetal medicine fellow and faculty physician (a routine practice at both study institutions). For purposes of data analysis the following occurrences were used as evidence of fetal compromise: (1) intrauterine fetal death, (2) positive backup test results (positive CST or biophysical profile score < 6), and (3) intervention because of an abnormal test. X2 or Fisher's exact tests were used for analysis of categoric variables. Student's t and Mann-Whitney rank-sum tests were used to compare interval and ordinal variables, respectively.
488
Weeks et al.
February 1995 Am J Obstet Gynecol
T a b l e II. Delivery indications by gestational age Gestational age at delivery
Causes of preterm delivery
<34 wk (n = 3)
34 to 35 wk 6 days (n = 18)
36 to 37 wk 6 days (n = 32)
38 to 39 wk 6 days (n = 125)
Positive test result Equivocal test result Labor Elective induction* Abruptio placentae Pregnancy-induced hypertension Otherf
1 (33) 2 (67) -
2 (11) 1 (5.6) 9 (50.0) 2 (11.1) 1 (5.6) 3 (16.7)
1 (3.1) 7 (21.9) 16 (50.0) 1 (3.1) 3 (9.4) 4 (12.5)
6 (4.8) I1 (8.8) 60 (48.0) 45 (36.0) 1 (0.8) 2 (1.6) -
>-40 wk (n = 122)
4 8 96 13
(3.3) (6.6) (78.7) (10.7) 1 (0,8) -
*Scheduled deliveries at >41 weeks were not considered elective. tPlacenta previa, patient anxiety, previous classic uterine scar.
T a b l e III. Pregnancy outcome by gestational age of previous stillbirth
EGA at delivery (wk, mean _+ SD) Birth weight (gm, mean -+- SD) EGA at first test (wk, mean _+ SD) Total tests (No., mean -+ SD) Abnormal test result (%)* Delivery for abnormal test result (%) C/S for fetal distress (%) Growth retardation (%)
Stillbirth <-32 wk (n = 115)
Stillbirth >-36 wk (n = 148)
38.9 +- 2.0 3270 -+ 560 32.1 + 4.0 13.9 -+ 10.7 61.7 20.0 3.5 5.2
39.1 -+ 2.0 3421 + 567 34.2 - 3.0 8.7 -+ 5.2 41.9 20.9 6.8 2.7
Significance
p p p p
= > > >
NS 0.03 0.001 0.001 0.01 NS NS NS
EGA, Estimated gestational age; NS, not significant; C/S, cesarean section.
*Positive or equivocal CST, biophysical profile score < 8.
Differences were considered to be significant when p < 0.05. These analyses were p e r f o r m e d with Epi Info 5 statistical software package. 13
Results T h e study group comprised 300 healthy p r e g n a n t w o m e n with a history of stillbirth as their only indication for a n t e p a r t u m testing. A probable cause for the previous stillbirth was identified in only 50% of cases. T h e most c o m m o n l y r e p o r t e d finding was "cord accident," accounting for nearly one third of stillbirths with a known cause of death. Other cited causes included abruptio placentae, pregnancy=induced hypertension, perinatal infection, a n d fetal anomalies (Table I). D u r i n g the study p e r i o d there was one r e c u r r e n t stillbirth a n d n o n e o n a t a l deaths, for a corrected perintal mortality rate of 3.3 p e r 1000. T h e intrauterine fetal death occurred in a 32-year-old, gravida 3, para 2, white w o m a n with a history of two u n e x p l a i n e d stillbirths. T h e prior stillbirths were at 38 a n d 37 weeks' gestation, respectively. D u r i n g h e r third p r e g n a n c y she was screened for chronic hypertension, diabetes, thyroid disease, sexually transmitted diseases, a n d collagen vascular disease, all of which were negative. T h e autopsy results from her two prior stillbirths had revealed structurally n o r m a l fetuses. Parental karyotypes were normal. T h e patient did well until 35 weeks' gestation,
when she was seen in the labor a n d delivery suite 3 d a y s after a negative CST c o m p l a i n i n g of decreased fetal movement. She was observed a n d discharged h o m e after a reactive N S T without decelerations. Approximately 16 hours later she r e t u r n e d c o m p l a i n i n g that she had perceived no fetal m o v e m e n t s for the preceding 5 hours. I n t r a u t e r i n e fetal death was confirmed, a n d she was subsequently delivered of an appropriately grown, structurally n o r m a l fetus. T h e r e was no placental pathologic m e c h a n i s m or other a p p a r e n t cause. Fifty-three of the 300 study patients (17.7%) were delivered at < 38 weeks' gestation. T h e most c o m m o n cause of early delivery was s p o n t a n e o u s labor, accounting for 51% of these deliveries. Equivocal or positive F H R tracings were the second most c o m m o n cause of early delivery, accounting for 23% (Table II). Overall, 13.6% of deliveries were effected after a b n o r m a l or positive test results. Fig. 1 shows the cumulative percentage of all patients with positive fetal test results a n d the cumulative percentage of all patients who were delivered for one or more positive tests at a given gestational age. Of the 300 patients tested, 19 (6.4%) had o n e or m o r e positive a n t e p a r t u m test results. Six patients (2% of the total study population) had the first positive test at < 36 weeks. T h r e e patients who had positive tests at 32 to 35 weeks' gestation were delivered, whereas three others
Volume 172, Number 2, Part 1 Am J Obstet Gynecol
Wesks ~t al. 489
Fig. 1. Cumulative percent of 19 patients with one or more positive test results. Fourteen patients were delivered because of positive test results.
with positive tests at < 32 weeks were followed up to term. None of the three patients with positive tests at < 32 weeks were delivered because of a positive test, and there were no cases of intrauterine growth retardation, cesarean section for fetal distress, or low Apgar scores among the three. Two patients initiated fetal surveillance at 26 weeks for a history of stillbirth at 30 and 32 weeks, respectively. Both patients were delivered at > 39 weeks, although their first positive tests occurred at 27 weeks. The third patient had her first positive test at 31 weeks, but labor was not induced until 37 weeks' gestation, when signs and symptoms of preeclampsia developed. Of the three patients who were delivered with positive tests at < 36 weeks, only one had unequivocal evidence of fetal compromise. That patient was seen at 34 weeks with a complaint of decreased fetal movement. She was found to have a 2/10 biophysical profile and was subsequently delivered by cesarean section for fetal distress. The two remaining patients had labor induction because of positive tests at 32 and 35 weeks' gestation and were eventually delivered of appropriately grown premature infants. To assess the impact, if any, of the gestational age at which the previous stillbirth occurred on the outcome of the subsequent pregnancy, we compared the patients in whom the prior stillbirth occurred at -< 32 weeks (n = 115) with those in whom the prior stillbirth occurred at -> 36 weeks (n = 148). T h e r e were no statistically significant differences in maternal age, gravidity, or parity. As anticipated, the group with a history of early stillbirths had significantly more tests p e r patient, a reflection of earlier antepartum testing (Table III). T h e group with early prior stillbirths also had significantly more abnormal tests; however, the incidences of intervention for an abnormal test, cesarean section for
tetal distress, and intrauterine growth retardation were similar between the two groups. Therefore it does not a p p e a r that a history of stillbirth at < 32 weeks is associated with poorer prognosis or worse outcome in the subsequent pregnancy. We then c o m p a r e d the obstetric outcomes of those patients whose cause of the previous stillbirth was known (n = 153) with those with unexplained previous stillbirths (n = 147). There were no statistically significant differences in maternal age, gravidity, or parity. We found a statistically significant decrease in mean birth weight among patients whose previous stillbirth was ur~explained (Table IV). The observed 160 gm difference was not clinically significant and was explained in part by the higher incidence of intrauterine growth retardation (7.5% vs 2.6%) among patients in the unknown group with unknown causes (p = 0.09). There were no differences in the frequency of abnormal tests, intervention for abnormal tests, or ceserean sections for fetal distress when siblings of stillborns with known versus unknown causes were c o m p a r e d (Table IV). It would a p p e a r that in an otherwise healthy pregnant woman, the cause of the previous stillbirth does not confer a negative or positive attribute on the outcome of subsequent pregnancies.
Comment In spite of the low incidence (3/1000) of recurrent stillbirth r e p o r t e d in this study we cannot exclude the possibility of a type II error. Against a background stillbirth rate of <1%, >3000 patients would be needed to evaluate the risk of recurrent stillbirth with 80% power. A definitive study that addresses the timing of fetal monitoring with recurrent stillbirth as an end point is seemingly unattainable; therefore we sought other indicators of increased fetal jeopardy. We defined
490 Weeks eta].
February 1995 AmJ Obstet Gynecol
Table IV, Pregnancy outcome by cause of previous stillbirth
EGA at delivery (wk, mean -+ SD) Birth weight (gm, mean _+ SD) EGA at first test (wk, mean + SD) Total tests (No., mean -+ SD) Abnormal test result (%)* Delivery for abnormal test result (%) Cesarean section for fetal distress (%) Fetal growth retardation (%)
Known (n = 147)
Unknown (n = 153)
39.1 -+ 1.9 3408 + 581 33.6 -+ 3.6 10.0 -+ 7.8 66 14.4 3.3 2.6
38.8 + 2.1 3244 -+ 543 32.8 -+ 3.7 12.1 -+ 5.7 61 13.0 6.8 7.5
[
Significance NS p = 0.01 NS NS NS NS NS p = 0.09
Abbreviations as in Table III. *Positive or equivocal CST, biophysical profile score < 8.
fetal compromise as stillbirth, positive antepartum tests, or intervention because of any abnormal test result. A significant percentage of the patients tested had at least one positive test (6.4%) or intervention because of an abnormal antepartum test result (13.6%). These data support what has been previously published, that such patients do constitute a high-risk group for whom antepartum surveillance is indicated? Indeed, our study patients were comparable to high-risk pregnancies complicated by hypertensive disorders or insulin-dep e n d e n t diabetes with regard to the incidence of abnormal tests and pregnancy interventions?' 6 The literature to date has not addressed when to initiate fetal testing in pregnancies complicated by a history of stillbirth. At our institution many clinicians have arbitrarily chosen to begin fetal monitoring 2 to 4 weeks before the gestational age of the previous stillbirth. If the previous stillbirth occurred at 34 weeks, is antepartum testing at 32 weeks any more rational or justified than testing at 30 weeks or conversely at 36 weeks? Although there are good data to guide us as to when to initiate fetal surveillance for mothers with hypertension, insulin-dependent diabetes, or collagen vascular disease, the same cannot be said for healthy mothers with previous stillbirths. Does knowing a cause of the previous stillbirth (e.g., cord accident) diminish the likelihood of untoward outcome in the subsequent pregnancy? If the cause of the prior stillbirth remains undetermined, does the patient have a poorer prognosis? None of these questions, which are frequently asked by affected patients, can be fully answered with our current knowledge of this clinical entity. This void generates anxiety in both the patient and her physician, which may result in fetal monitoring at very premature gestations or even preterm pregnancy termination, in spite of the lack of data to support such an approach. In our study of 300 "healthy" pregnant women with a history of stillbirth, the earliest positive tests were at 27, 27, and 31 weeks, respectively. However, these tests were false positive, because all of these mothers were
delivered at term for indications other than fetal distress. We were unable to demonstrate a relationship between the gestational age of the previous stillbirth and future pregnancy outcome. In our study population patients who had previous stillbirths at < 32 weeks did not appear to have a higher incidence of fetal compromise compared with those with previous stillbirths at -> 36 weeks. Furthermore, the earliest delivery in association With a positive test result occurred at 32 weeks. After our initial review of these data, it was determined that ascertainment bias could have resulted in our failure to show earlier fetal compromise in patients who had a previous stillbirth at -< 32 weeks. Only patients who were seen for antepartum surveillance were included in the study. Consequently, recurrent stillbirths among patients who had not yet begun antepartum testing could have been missed. We therefore reviewed our perinatal database for fetal deaths and found no patients who would have met our inclusion criteria (i.e., there were no recurrent stillbirths in the absence of maternal medical complications). In constructing an algorithm for antepartum surveillance in mothers with high-risk conditions, there are a n u m b e r of important considerations. Although ideally testing should be initiated before evidence of fetal compromise, it is often impossible to pinpoint the onset of fetal peril. In fact, our one fetal death in spite of normal antepartum testing suggests that in some instances fetal death cannot be predicted with our most valued monitoring techniques. Another consideration is that antenatal fetal surveillance is highly gestational-age dependent. At 28 weeks' gesatation as few as 60% of fetuses will have NSTs that meet the accepted criteria for reactivity. 14 This is not a result of uteroplacental insufficiency but rather the reflection of an immature fetal autonomic nervous system. Therefore monitoring of the viable yet very premature fetus frequently requires more elaborate and expensive backup testing. Undoubtedly this also engenders a high degree of anxiety in the patient who has already had a loss in a
Volume 172, Number 2, Part 1
W e e k s et al,
491
Anl J Obstet Gynecol
Maternal H&P Review details of Previous Stillbirth No medical complications
fetal "kick counts" at ~ 26-28 wks
\
chronic medical conditions or (+) Lupus screen
decreased FM
BPP's with selective use of NST at 26-28wks (16)
normal fetal movement
/
semi-weekly MBPP or
weekly BPP or CST at 32 wks
normal testing
continue to term
persistent (+) CST or BPP
delivery
frequent abnormal tests
assess fetal maturity & cervix
Fig. 2. Proposed antepartum testing scheme for patients with history of stillbirth. H & P, History and physical examination; FM, fetal movements; BPP, biophysical profile; MBPP, modified biophysical profile.
previous pregnancy. On the basis of these observations and the results o f our study, we offer the following guidelines for antepartum surveillance in the otherwise healthy pregnant woman with a history of previous stillbirth (Fig. 2). We propose that electronic FHR monitoring be started at 32 weeks or more, a gestational age when virtually all fetuses have reactivity. At 26 to 28 weeks' gestation all patients with prior stillbirth should be instructed to begin fetal movement assessment as described by Moore and Piacquadio. 15 Patients who report decreased fetal movement should have follow-up fetal surveillance. Finally, for those patients and clinicians whose anxiety prompts antepartum testing at viability, biophysical assessment with selective use of the NST should minimize the number of backup tests performed after nonreactive or equivocal tests. 16Adherence to these recommendations will undoubtedly decrease the incidence of abnormal tests resulting from fetal immaturity rather than true uteroplacental insufficiency.
This protocol is not appropriate for all patients with a history of stillbirth. Some fetal deaths are the result of nonrecurring conditions such as perinatal infection, fetal anomalies, or maternal trauma. Affected mothers would be better served by punctilious prenatal screening, education, and counseling in future pregnancies. Conversely, some stillbirths follow obstetric complications that can recur but cannot be predicted (e.g., abruptio placentae, cord prolapse, uterine rupture). Clearly, these are patients who are unlikely to benefit from antepartum surveillance in subsequent pregnancies. Some may question why we recommend antepartum surveillance at all, noting that (1) we had one recurrent stillbirth in spite of normal antepartum testing (a case in which the mother's perception of decreased fetal movement proved to be a more reliable indicator of fetal health than did electronic fetal monitoring), (2) the only patient with clear evidence of fetal compromise after delivery for a positive test at < 3 6 weeks also had decreased fetal movement, and (3) when
492
Weeks et al.
the 3.3 p e r 1000 perinatal mortality rate of our study group is c o m p a r e d with the c u r r e n t 7.5 per 1000 rate nationwide, ~7 healthy patients with a history of previous stillbirth do n o t a p p e a r to be at increased risk for fetal death. Admittedly, these data b r i n g to question the importance of previous stillbirth as a risk factor in future pregnancies a n d the relative value of electronic fetal m o n i t o r i n g over m a t e r n a l assessment of fetal movement. However, the low perinatal mortality rate in our study may have b e e n affected by a n t e p a r t u m testing with high rates of intervention before the d e v e l o p m e n t of overt or severe fetal compromise. Furthermore, we c a n n o t dismiss the possibility of a higher perinatal mortality risk that was u n d e t e c t e d because of inadequate power. Finally, with regard to fetal m o v e m e n t assessment, we believe that "fetal kick counting" is a n effective screening technique. But, until a n adequately sized, prospective study proves its ability to s u p p l a n t F H R a n d u l t r a s o n o g r a p h i c monitoring, we will continue to r e c o m m e n d the latter forms of a n t e p a r t u m surveillance for patients with prior stillbirths. A history of stillbirth is widely recognized as an indication for a n t e p a r t u m fetal surveillance. T h e p r o p e r time to initiate m o n i t o r i n g has n o t b e e n previously studied. In our study group of 300 healthy pregn a n t w o m e n with a history of stillbirth as the only indication for testing, there were no p r e g n a n c y interventions for a b n o r m a l tests before 32 weeks' gestation. Moreover, we f o u n d n o differences in future p r e g n a n c y outcome between patients who had a p r e t e r m stillbirth a n d those who had a previous term fetal death. We therefore propose initiating a n t e p a r t u m surveillance at 32 weeks or later. Although this r e c o m m e n d a t i o n to p o s t p o n e fetal m o n i t o r i n g until 32 weeks may miss a rare patient with earlier fetal compromise, we believe that this scheme of surveillance will significantly diminish the n e e d for repeated backup testing a n d the a t t e n d a n t increase in patient a n d physician anxiety. REFERENCES
1. Butler M, Bongam D. Perinatal mortality: the first report of the British Perinatal Mortality Survey. Edinburgh: Livinston, 1963:32-7.
February 1995 AmJ Obstet Gynecol
2. Niswander K, Gordon M. Collaborative prenatal study of the National Institute of Neurologic Disease and Stroke: the women and their pregnancies. Bethesda, Maryland: National Institutes of Health, 1972; DHEW publication no (NIH) 73-379. 3. Freeman R, Dorchester W, Anderson G, Garite T. The significance of a previous stillbirth. AM J OBSTETGVNECOL 1985;151:7-13. 4. Freeman R, Garite T, Nageotte M. Fetal heart rate monitoring. 2nd ed. Baltimore: Williams & Wilkins, 1991:158. 5. Lagrew D, Pircon R, Towers C, Dorchester W, Freeman R. Antepartum fetal surveillance in patients with diabetes: when to start. AMJ OBSTETGYNECOL1993;168:1820-6. 6. Pircon R, Lagrew D, Towers C, Dorchester W, Gocke S, Freeman R. Antepartum testing in the hypertensive patient: when to begin. AM J OBSTETGYNECOL1991;164: 1563-70. 7. Williams R, Creasy R, Cunningham G, Hawes W, Norris F, Tashiro M. Fetal growth and perinatal viability in California. Obstet Gynecol 1982;59:624-32. 8. Eammer E, Brown L, Anderka M, Guyer B. Classification and analysis of fetal deaths in Massachusetts. JAMA 1989; 261:1757-62. 9. Morrison I, Olsen J. Weight-specific stillbirths and associated causes of death: an analysis of 765 stillbirths. 1985; 152:975-80. 10. Eden R, Seifert L, Kodack L. A modified biophysical profile for antenatal fetal surveillance. Obstet Gynecol 1988;71:365. 11. Clark S, Sabey P, Jolley D. Nonstress testing with acoustic stimulation and amniotic fluid volume assessment: 5973 tests without unexpected fetal death. AMJ OBSTETGYNECOL 1989;160:694-7. 12. Phelan J, Smith C, Broussard P, Small M. Amniotic fluid volume assessment with the four-quadrant technique at 36-42 weeks gestation. J Reprod Med 1987;32:540-2. 13. Dean A, Dean J, Burton A, Dicker R. Epi Info, version 5: a word processing, database, and statistics program for epidemiology on microcomputers. Stone Mountain, Georgia: USD, Inc, 1990. 14. Carlan S, Gore M, VanMeter S, Mastrogiannis D. A longitudinal study evaluating the effect of gestational age on antenatal assessment tests. AM J OBSTETGYNECOL 1992; 166(suppl):412. 15. Moore T, Piacquadio K. A prospective evaluation of fetal movement screening to reduce the incidence of anteparturn fetal death. 1989;160:1075-80. 16. Manning F, Morrison I, Lange I, Harman C. Fetal biophysical profile scoring: selective use of the nonstress test. 1987;156:709-12. 17. Cunningham F, MacDonald E Gant N, Leveno K, Gilstrap L. Williams' obstetrics. 19th ed. Norwalk, Connecticut: Appleton & Lange, 1993:5-6.
Weeks eta[.
20. Fine RN. Diagnosis and treatment of fetal urinary tract abnormalities. J Pediatr 1992; 121:333-41. 21. Roberts JA, Kaack MB, Morvant AB. Vesicoureteral reflux in the primate. IV. Infection as cause of prolonged highgrade reflux. Pediatrics 1988;82:91-5. 22. Najmaldin A, Burge DM, Atwell JD. Reflux nephropathy secondary to intrauterine vesicoureteric reflux. J Pediatr Surg 1990;25:387-90. 23. Steele BT, Robitaille P, De Maria J, Grignon A. Follow-up
evaluation of prenatally recognised vesicoureteric reflux. J Pediatr 1989;115;95-6. 24. Gordon AC, Thomas DFM, Arthur RJ, Irving HC, Smith SEW. Prenatally diagnosed reflux: a follow-up study. Br J Urol 1990;65:407-12. 25. Zerin JM, Ritchey ML, Chang ACH. Incidental vesicoureteral reflux in neonates with antenatally detected hydronephrosis and other renal abnormalities. Radiology 1993;187:157-60.
Antepartum surveillance for a history of stillbirth: When to begin? Jonathan W. Weeks, MD," Tamerou Asrat, MD, b Mark A. Morgan, MD, b Michael Nageotte, MD, b Steven J. Thomas, MD, c and Roger K. Freeman, MD b Louisville, Kentucky, and Irvine, California OBJECTIVE: A history of stillbirth is universally accepted as an indication for antepartum fetal heart rate testing. Our goal was to examine when fetal testing should begin in an otherwise healthy patient with a history of stillbirth. STUDY DESIGN: This is a nonconcurrent cohort study of patients who were seen for antepartum surveillance from January 1979 to December 1991 with a history of stillbirth as the only indication for testing. Subsequent pregnancies were evaluated for adverse outcomes and abnormal antepartum test results. RESULTS: There was one case of recurrent stillbirth among the 300 study patients. Nineteen patients (6.4%) had one or more positive antepartum surveillance tests (positive contraction stress test or biophysical profile -<4). Three patients (1%) had positive tests before 32 weeks, all of whom were subsequently delivered without incident at term. Three patients were delivered for positive tests at < 36 weeks, one by cesarean section for fetal distress. We could not detect a relationship between the gestational age of the previous stillborn and the incidence of abnormal tests or fetal distress in subsequent pregnancies. CONCLUSION: Antepartum surveillance should begin at ->32 weeks in the healthy pregnant woman with a history of stillbirth. (AM J OBSTETGYNECOL1995;172:486-92.)
Key words: Stillbirth, antepartum surveillance, fetal heart rate testing, fetal monitoring
A history of stillbirth has long been recognized as a significant risk factor for subsequent poor pregnancy outcome. The perinatal collaborative studies in the United States and Britain demonstrated an increased risk of perinatal mortality and neurologic damage in subsequent births?' 2 Since the completion of these landmark studies there has been only one United States From the Department of Obstetrics and Gynecology, University of Louisville, a and the Department of Obstetrics and Gynecologyband the Division of Perinatology/ University of California, Irvine. Received for publication January 26, 1994; revised April 12, 1994; acceptedJuly 21, 1994. Reprint requests:Jonathan W. Weeks, MD, Department of Obstetrics and Gynecology, 2nd floor Ambulatory Care Bldg, 550 S. Jackson St., Louisville, KY 40292. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/59397 486
study to evaluate the significance of a history of stillbirth in the era of electronic fetal surveillance. Freeman et al? demonstrated that a history of stillbirth remains a significant risk factor and concluded that such patients deserve antepartum fetal testing. However, the p r o p e r time to initiate fetal monitoring remains undetermined. Although some sources r e c o m m e n d beginning at 32 to 34 weeks, 4 many clinicians initiate monitoring 2 to 4 weeks before the gestational age of the previous stillbirth. In part, this controversy stems from the paucity of published data on the timing of fetal monitoring in patients with a previous stillbirth. Approximately half of all stillbirths are the result of significant maternal medical conditions, the most common being hypertensive disorders and insulin-dependent diabetes. When to initiate antepartum surveillance
Volume 172, Number 2, Part 1 AmJ Obstet Gynecol
in such patients has been studied and r e p o r t e d in two prior publications from this institutionJ' 6 We now report our experience with antepartum surveillance in the otherwise healthy patient with a previous stillbirth. These patients were evaluated for adverse pregnancy outcome and antepartum tests results indicative of uteroplacental insufficiency. Our goal was to determine when to initiate antepartum fetal surveillance in healthy patients with a history of stillbirth. Material and methods All study patients were seen in the Centers for Fetal Evaluation at Long Beach Memorial and the University of California, Irvine, medical centers. The Long Beach Memorial antepartum testing computer database was reviewed from January 1979 through December 1991. A similar database at the University Of California, Irvine, was analyzed for patients tested from January 1988 through December 1991. During this period approximately 70,000 tests were performed on > 15,000 patients. The databases were queried for patients who were referred with singleton pregnancies and a history ,, of stillbirth as the only indication for testing. Medical records were reviewed to confirm the accuracy of the coded referral diagnosis. Patients with a history of intrauterine fetal death at < 20 weeks' gestation were excluded. The search yielded 39 patients from the University of California, Irvine, and 261 from Long Beach Memorial Medical Center. Approximately 15 patients from the Long Beach Memorial database had been r e p o r t e d in a previous collaborative study on antepartum fetal heart rate (FHR) testing? Patient histories, antepartum fetal test results, and pregnancy outcomes were prospectively collected and retrospectively analyzed. Recorded patient history included maternal age, gravidity, parity, gestational age when tested, testing method, test results, gestational age at delivery, indication for delivery, other medical or obstetric diagnoses, neonatal Apgar scores, neonatal intensive care admission, and birth weight. Neonates with birth weights < 10th percentile by California growth chart standards were considered to be growth restricted. 7 Obstetric or medical complications associated with the previous stillbirth were ascertained by reviewing the medical records. We placed an emphasis on describing the clinical setting in which the previous stillbirths occurred rather than relying on the death certificate and autopsy-derived classification systems that have been previously described. 8' 9 Autopsy reports were used in cases where the stillborn infants were delivered at the study institutions. When the cause of the previous stillbirth was stated to be an event that resulted in umbilical cord occlusion (e.g., prolapse, torsion, thrombosis, multiple tight loops, true knots), it was classified as a "cord accident." Pregnancy-induced hypertension was the assigned cause when an otherwise unexplained,
Weeks et al. 487
Table I. Causes of previous stillbirth Unknown Cord accident Infection Anomalies Abruptio placentae PIH Postdates Twins Trauma Intrapartum Drugs or alcohol IUGR Placental infarct Uterine rupture Other
147 48 13 13 14 12 10 9 6 6 5 5 4 1 7
(49.0) (16.0) (4.3) (4.3) (4.6) (4.0) (3.3) (3.0) (2.0) (2.0) (1.6) (1.6) (1.3) (0.3) (2.3)
PIH, Pregnancy-induced hypertension; IUGR, intrauterine growth retardation.
structurally normal stillborn infant occurred in a mother with signs and symptoms of pregnancy-induced hypertension or preeclampsia. Infection was considered to be the cause when there was funisitis or pneumonia on histopathologic study documented perinatal infection (e.g., syphilis or listeria), or when the mother was admitted with classic signs and symptoms of chorioamnionitis. Up until December 1989 the primary method of surveillance was weekly contraction stress tests (CST). Testing procedures and test interpretation have been described elsewhere. 4 From January 1990 through December 1991 the primary mode of surveillance was changed to a semiweekly modified biophysical profile (nonstress test [NST] and amniotic fluid index)J °' ~1 Tests were read as reactive if there were two accelerations that reached a peak of 15 beats/min above the baseline and lasted 15 seconds in a 20-minute window of recording time. The amniotic fluid index was performed as described by Phelan et alJ 2 If the NST was nonreactive or had FHR deceleration or if the amniotic fluid index was <5.0 cm, a CST or full biophysical profile was performed as "backup" for the abnormal primary test. Backup tests were performed in all instances of abnormal primary tests even if immediate delivery was planned. Intervention because of an abnormal test result was the decision of the primary physician, often in consultation with the perinatal staff. All FHR tests were prospectively read by at least one maternal-fetal medicine fellow and faculty physician (a routine practice at both study institutions). For purposes of data analysis the following occurrences were used as evidence of fetal compromise: (1) intrauterine fetal death, (2) positive backup test results (positive CST or biophysical profile score < 6), and (3) intervention because of an abnormal test. X2 or Fisher's exact tests were used for analysis of categoric variables. Student's t and Mann-Whitney rank-sum tests were used to compare interval and ordinal variables, respectively.
488
Weeks et al.
February 1995 Am J Obstet Gynecol
T a b l e II. Delivery indications by gestational age Gestational age at delivery
Causes of preterm delivery
<34 wk (n = 3)
34 to 35 wk 6 days (n = 18)
36 to 37 wk 6 days (n = 32)
38 to 39 wk 6 days (n = 125)
Positive test result Equivocal test result Labor Elective induction* Abruptio placentae Pregnancy-induced hypertension Otherf
1 (33) 2 (67) -
2 (11) 1 (5.6) 9 (50.0) 2 (11.1) 1 (5.6) 3 (16.7)
1 (3.1) 7 (21.9) 16 (50.0) 1 (3.1) 3 (9.4) 4 (12.5)
6 (4.8) I1 (8.8) 60 (48.0) 45 (36.0) 1 (0.8) 2 (1.6) -
>-40 wk (n = 122)
4 8 96 13
(3.3) (6.6) (78.7) (10.7) 1 (0,8) -
*Scheduled deliveries at >41 weeks were not considered elective. tPlacenta previa, patient anxiety, previous classic uterine scar.
T a b l e III. Pregnancy outcome by gestational age of previous stillbirth
EGA at delivery (wk, mean _+ SD) Birth weight (gm, mean -+- SD) EGA at first test (wk, mean _+ SD) Total tests (No., mean -+ SD) Abnormal test result (%)* Delivery for abnormal test result (%) C/S for fetal distress (%) Growth retardation (%)
Stillbirth <-32 wk (n = 115)
Stillbirth >-36 wk (n = 148)
38.9 +- 2.0 3270 -+ 560 32.1 + 4.0 13.9 -+ 10.7 61.7 20.0 3.5 5.2
39.1 -+ 2.0 3421 + 567 34.2 - 3.0 8.7 -+ 5.2 41.9 20.9 6.8 2.7
Significance
p p p p
= > > >
NS 0.03 0.001 0.001 0.01 NS NS NS
EGA, Estimated gestational age; NS, not significant; C/S, cesarean section.
*Positive or equivocal CST, biophysical profile score < 8.
Differences were considered to be significant when p < 0.05. These analyses were p e r f o r m e d with Epi Info 5 statistical software package. 13
Results T h e study group comprised 300 healthy p r e g n a n t w o m e n with a history of stillbirth as their only indication for a n t e p a r t u m testing. A probable cause for the previous stillbirth was identified in only 50% of cases. T h e most c o m m o n l y r e p o r t e d finding was "cord accident," accounting for nearly one third of stillbirths with a known cause of death. Other cited causes included abruptio placentae, pregnancy=induced hypertension, perinatal infection, a n d fetal anomalies (Table I). D u r i n g the study p e r i o d there was one r e c u r r e n t stillbirth a n d n o n e o n a t a l deaths, for a corrected perintal mortality rate of 3.3 p e r 1000. T h e intrauterine fetal death occurred in a 32-year-old, gravida 3, para 2, white w o m a n with a history of two u n e x p l a i n e d stillbirths. T h e prior stillbirths were at 38 a n d 37 weeks' gestation, respectively. D u r i n g h e r third p r e g n a n c y she was screened for chronic hypertension, diabetes, thyroid disease, sexually transmitted diseases, a n d collagen vascular disease, all of which were negative. T h e autopsy results from her two prior stillbirths had revealed structurally n o r m a l fetuses. Parental karyotypes were normal. T h e patient did well until 35 weeks' gestation,
when she was seen in the labor a n d delivery suite 3 d a y s after a negative CST c o m p l a i n i n g of decreased fetal movement. She was observed a n d discharged h o m e after a reactive N S T without decelerations. Approximately 16 hours later she r e t u r n e d c o m p l a i n i n g that she had perceived no fetal m o v e m e n t s for the preceding 5 hours. I n t r a u t e r i n e fetal death was confirmed, a n d she was subsequently delivered of an appropriately grown, structurally n o r m a l fetus. T h e r e was no placental pathologic m e c h a n i s m or other a p p a r e n t cause. Fifty-three of the 300 study patients (17.7%) were delivered at < 38 weeks' gestation. T h e most c o m m o n cause of early delivery was s p o n t a n e o u s labor, accounting for 51% of these deliveries. Equivocal or positive F H R tracings were the second most c o m m o n cause of early delivery, accounting for 23% (Table II). Overall, 13.6% of deliveries were effected after a b n o r m a l or positive test results. Fig. 1 shows the cumulative percentage of all patients with positive fetal test results a n d the cumulative percentage of all patients who were delivered for one or more positive tests at a given gestational age. Of the 300 patients tested, 19 (6.4%) had o n e or m o r e positive a n t e p a r t u m test results. Six patients (2% of the total study population) had the first positive test at < 36 weeks. T h r e e patients who had positive tests at 32 to 35 weeks' gestation were delivered, whereas three others
Volume 172, Number 2, Part 1 Am J Obstet Gynecol
Wesks ~t al. 489
Fig. 1. Cumulative percent of 19 patients with one or more positive test results. Fourteen patients were delivered because of positive test results.
with positive tests at < 32 weeks were followed up to term. None of the three patients with positive tests at < 32 weeks were delivered because of a positive test, and there were no cases of intrauterine growth retardation, cesarean section for fetal distress, or low Apgar scores among the three. Two patients initiated fetal surveillance at 26 weeks for a history of stillbirth at 30 and 32 weeks, respectively. Both patients were delivered at > 39 weeks, although their first positive tests occurred at 27 weeks. The third patient had her first positive test at 31 weeks, but labor was not induced until 37 weeks' gestation, when signs and symptoms of preeclampsia developed. Of the three patients who were delivered with positive tests at < 36 weeks, only one had unequivocal evidence of fetal compromise. That patient was seen at 34 weeks with a complaint of decreased fetal movement. She was found to have a 2/10 biophysical profile and was subsequently delivered by cesarean section for fetal distress. The two remaining patients had labor induction because of positive tests at 32 and 35 weeks' gestation and were eventually delivered of appropriately grown premature infants. To assess the impact, if any, of the gestational age at which the previous stillbirth occurred on the outcome of the subsequent pregnancy, we compared the patients in whom the prior stillbirth occurred at -< 32 weeks (n = 115) with those in whom the prior stillbirth occurred at -> 36 weeks (n = 148). T h e r e were no statistically significant differences in maternal age, gravidity, or parity. As anticipated, the group with a history of early stillbirths had significantly more tests p e r patient, a reflection of earlier antepartum testing (Table III). T h e group with early prior stillbirths also had significantly more abnormal tests; however, the incidences of intervention for an abnormal test, cesarean section for
tetal distress, and intrauterine growth retardation were similar between the two groups. Therefore it does not a p p e a r that a history of stillbirth at < 32 weeks is associated with poorer prognosis or worse outcome in the subsequent pregnancy. We then c o m p a r e d the obstetric outcomes of those patients whose cause of the previous stillbirth was known (n = 153) with those with unexplained previous stillbirths (n = 147). There were no statistically significant differences in maternal age, gravidity, or parity. We found a statistically significant decrease in mean birth weight among patients whose previous stillbirth was ur~explained (Table IV). The observed 160 gm difference was not clinically significant and was explained in part by the higher incidence of intrauterine growth retardation (7.5% vs 2.6%) among patients in the unknown group with unknown causes (p = 0.09). There were no differences in the frequency of abnormal tests, intervention for abnormal tests, or ceserean sections for fetal distress when siblings of stillborns with known versus unknown causes were c o m p a r e d (Table IV). It would a p p e a r that in an otherwise healthy pregnant woman, the cause of the previous stillbirth does not confer a negative or positive attribute on the outcome of subsequent pregnancies.
Comment In spite of the low incidence (3/1000) of recurrent stillbirth r e p o r t e d in this study we cannot exclude the possibility of a type II error. Against a background stillbirth rate of <1%, >3000 patients would be needed to evaluate the risk of recurrent stillbirth with 80% power. A definitive study that addresses the timing of fetal monitoring with recurrent stillbirth as an end point is seemingly unattainable; therefore we sought other indicators of increased fetal jeopardy. We defined
490 Weeks eta].
February 1995 AmJ Obstet Gynecol
Table IV, Pregnancy outcome by cause of previous stillbirth
EGA at delivery (wk, mean -+ SD) Birth weight (gm, mean _+ SD) EGA at first test (wk, mean + SD) Total tests (No., mean -+ SD) Abnormal test result (%)* Delivery for abnormal test result (%) Cesarean section for fetal distress (%) Fetal growth retardation (%)
Known (n = 147)
Unknown (n = 153)
39.1 -+ 1.9 3408 + 581 33.6 -+ 3.6 10.0 -+ 7.8 66 14.4 3.3 2.6
38.8 + 2.1 3244 -+ 543 32.8 -+ 3.7 12.1 -+ 5.7 61 13.0 6.8 7.5
[
Significance NS p = 0.01 NS NS NS NS NS p = 0.09
Abbreviations as in Table III. *Positive or equivocal CST, biophysical profile score < 8.
fetal compromise as stillbirth, positive antepartum tests, or intervention because of any abnormal test result. A significant percentage of the patients tested had at least one positive test (6.4%) or intervention because of an abnormal antepartum test result (13.6%). These data support what has been previously published, that such patients do constitute a high-risk group for whom antepartum surveillance is indicated? Indeed, our study patients were comparable to high-risk pregnancies complicated by hypertensive disorders or insulin-dep e n d e n t diabetes with regard to the incidence of abnormal tests and pregnancy interventions?' 6 The literature to date has not addressed when to initiate fetal testing in pregnancies complicated by a history of stillbirth. At our institution many clinicians have arbitrarily chosen to begin fetal monitoring 2 to 4 weeks before the gestational age of the previous stillbirth. If the previous stillbirth occurred at 34 weeks, is antepartum testing at 32 weeks any more rational or justified than testing at 30 weeks or conversely at 36 weeks? Although there are good data to guide us as to when to initiate fetal surveillance for mothers with hypertension, insulin-dependent diabetes, or collagen vascular disease, the same cannot be said for healthy mothers with previous stillbirths. Does knowing a cause of the previous stillbirth (e.g., cord accident) diminish the likelihood of untoward outcome in the subsequent pregnancy? If the cause of the prior stillbirth remains undetermined, does the patient have a poorer prognosis? None of these questions, which are frequently asked by affected patients, can be fully answered with our current knowledge of this clinical entity. This void generates anxiety in both the patient and her physician, which may result in fetal monitoring at very premature gestations or even preterm pregnancy termination, in spite of the lack of data to support such an approach. In our study of 300 "healthy" pregnant women with a history of stillbirth, the earliest positive tests were at 27, 27, and 31 weeks, respectively. However, these tests were false positive, because all of these mothers were
delivered at term for indications other than fetal distress. We were unable to demonstrate a relationship between the gestational age of the previous stillbirth and future pregnancy outcome. In our study population patients who had previous stillbirths at < 32 weeks did not appear to have a higher incidence of fetal compromise compared with those with previous stillbirths at -> 36 weeks. Furthermore, the earliest delivery in association With a positive test result occurred at 32 weeks. After our initial review of these data, it was determined that ascertainment bias could have resulted in our failure to show earlier fetal compromise in patients who had a previous stillbirth at -< 32 weeks. Only patients who were seen for antepartum surveillance were included in the study. Consequently, recurrent stillbirths among patients who had not yet begun antepartum testing could have been missed. We therefore reviewed our perinatal database for fetal deaths and found no patients who would have met our inclusion criteria (i.e., there were no recurrent stillbirths in the absence of maternal medical complications). In constructing an algorithm for antepartum surveillance in mothers with high-risk conditions, there are a n u m b e r of important considerations. Although ideally testing should be initiated before evidence of fetal compromise, it is often impossible to pinpoint the onset of fetal peril. In fact, our one fetal death in spite of normal antepartum testing suggests that in some instances fetal death cannot be predicted with our most valued monitoring techniques. Another consideration is that antenatal fetal surveillance is highly gestational-age dependent. At 28 weeks' gesatation as few as 60% of fetuses will have NSTs that meet the accepted criteria for reactivity. 14 This is not a result of uteroplacental insufficiency but rather the reflection of an immature fetal autonomic nervous system. Therefore monitoring of the viable yet very premature fetus frequently requires more elaborate and expensive backup testing. Undoubtedly this also engenders a high degree of anxiety in the patient who has already had a loss in a
Volume 172, Number 2, Part 1
W e e k s et al,
491
Anl J Obstet Gynecol
Maternal H&P Review details of Previous Stillbirth No medical complications
fetal "kick counts" at ~ 26-28 wks
\
chronic medical conditions or (+) Lupus screen
decreased FM
BPP's with selective use of NST at 26-28wks (16)
normal fetal movement
/
semi-weekly MBPP or
weekly BPP or CST at 32 wks
normal testing
continue to term
persistent (+) CST or BPP
delivery
frequent abnormal tests
assess fetal maturity & cervix
Fig. 2. Proposed antepartum testing scheme for patients with history of stillbirth. H & P, History and physical examination; FM, fetal movements; BPP, biophysical profile; MBPP, modified biophysical profile.
previous pregnancy. On the basis of these observations and the results o f our study, we offer the following guidelines for antepartum surveillance in the otherwise healthy pregnant woman with a history of previous stillbirth (Fig. 2). We propose that electronic FHR monitoring be started at 32 weeks or more, a gestational age when virtually all fetuses have reactivity. At 26 to 28 weeks' gestation all patients with prior stillbirth should be instructed to begin fetal movement assessment as described by Moore and Piacquadio. 15 Patients who report decreased fetal movement should have follow-up fetal surveillance. Finally, for those patients and clinicians whose anxiety prompts antepartum testing at viability, biophysical assessment with selective use of the NST should minimize the number of backup tests performed after nonreactive or equivocal tests. 16Adherence to these recommendations will undoubtedly decrease the incidence of abnormal tests resulting from fetal immaturity rather than true uteroplacental insufficiency.
This protocol is not appropriate for all patients with a history of stillbirth. Some fetal deaths are the result of nonrecurring conditions such as perinatal infection, fetal anomalies, or maternal trauma. Affected mothers would be better served by punctilious prenatal screening, education, and counseling in future pregnancies. Conversely, some stillbirths follow obstetric complications that can recur but cannot be predicted (e.g., abruptio placentae, cord prolapse, uterine rupture). Clearly, these are patients who are unlikely to benefit from antepartum surveillance in subsequent pregnancies. Some may question why we recommend antepartum surveillance at all, noting that (1) we had one recurrent stillbirth in spite of normal antepartum testing (a case in which the mother's perception of decreased fetal movement proved to be a more reliable indicator of fetal health than did electronic fetal monitoring), (2) the only patient with clear evidence of fetal compromise after delivery for a positive test at < 3 6 weeks also had decreased fetal movement, and (3) when
492
Weeks et al.
the 3.3 p e r 1000 perinatal mortality rate of our study group is c o m p a r e d with the c u r r e n t 7.5 per 1000 rate nationwide, ~7 healthy patients with a history of previous stillbirth do n o t a p p e a r to be at increased risk for fetal death. Admittedly, these data b r i n g to question the importance of previous stillbirth as a risk factor in future pregnancies a n d the relative value of electronic fetal m o n i t o r i n g over m a t e r n a l assessment of fetal movement. However, the low perinatal mortality rate in our study may have b e e n affected by a n t e p a r t u m testing with high rates of intervention before the d e v e l o p m e n t of overt or severe fetal compromise. Furthermore, we c a n n o t dismiss the possibility of a higher perinatal mortality risk that was u n d e t e c t e d because of inadequate power. Finally, with regard to fetal m o v e m e n t assessment, we believe that "fetal kick counting" is a n effective screening technique. But, until a n adequately sized, prospective study proves its ability to s u p p l a n t F H R a n d u l t r a s o n o g r a p h i c monitoring, we will continue to r e c o m m e n d the latter forms of a n t e p a r t u m surveillance for patients with prior stillbirths. A history of stillbirth is widely recognized as an indication for a n t e p a r t u m fetal surveillance. T h e p r o p e r time to initiate m o n i t o r i n g has n o t b e e n previously studied. In our study group of 300 healthy pregn a n t w o m e n with a history of stillbirth as the only indication for testing, there were no p r e g n a n c y interventions for a b n o r m a l tests before 32 weeks' gestation. Moreover, we f o u n d n o differences in future p r e g n a n c y outcome between patients who had a p r e t e r m stillbirth a n d those who had a previous term fetal death. We therefore propose initiating a n t e p a r t u m surveillance at 32 weeks or later. Although this r e c o m m e n d a t i o n to p o s t p o n e fetal m o n i t o r i n g until 32 weeks may miss a rare patient with earlier fetal compromise, we believe that this scheme of surveillance will significantly diminish the n e e d for repeated backup testing a n d the a t t e n d a n t increase in patient a n d physician anxiety. REFERENCES
1. Butler M, Bongam D. Perinatal mortality: the first report of the British Perinatal Mortality Survey. Edinburgh: Livinston, 1963:32-7.
February 1995 AmJ Obstet Gynecol
2. Niswander K, Gordon M. Collaborative prenatal study of the National Institute of Neurologic Disease and Stroke: the women and their pregnancies. Bethesda, Maryland: National Institutes of Health, 1972; DHEW publication no (NIH) 73-379. 3. Freeman R, Dorchester W, Anderson G, Garite T. The significance of a previous stillbirth. AM J OBSTETGVNECOL 1985;151:7-13. 4. Freeman R, Garite T, Nageotte M. Fetal heart rate monitoring. 2nd ed. Baltimore: Williams & Wilkins, 1991:158. 5. Lagrew D, Pircon R, Towers C, Dorchester W, Freeman R. Antepartum fetal surveillance in patients with diabetes: when to start. AMJ OBSTETGYNECOL1993;168:1820-6. 6. Pircon R, Lagrew D, Towers C, Dorchester W, Gocke S, Freeman R. Antepartum testing in the hypertensive patient: when to begin. AM J OBSTETGYNECOL1991;164: 1563-70. 7. Williams R, Creasy R, Cunningham G, Hawes W, Norris F, Tashiro M. Fetal growth and perinatal viability in California. Obstet Gynecol 1982;59:624-32. 8. Eammer E, Brown L, Anderka M, Guyer B. Classification and analysis of fetal deaths in Massachusetts. JAMA 1989; 261:1757-62. 9. Morrison I, Olsen J. Weight-specific stillbirths and associated causes of death: an analysis of 765 stillbirths. 1985; 152:975-80. 10. Eden R, Seifert L, Kodack L. A modified biophysical profile for antenatal fetal surveillance. Obstet Gynecol 1988;71:365. 11. Clark S, Sabey P, Jolley D. Nonstress testing with acoustic stimulation and amniotic fluid volume assessment: 5973 tests without unexpected fetal death. AMJ OBSTETGYNECOL 1989;160:694-7. 12. Phelan J, Smith C, Broussard P, Small M. Amniotic fluid volume assessment with the four-quadrant technique at 36-42 weeks gestation. J Reprod Med 1987;32:540-2. 13. Dean A, Dean J, Burton A, Dicker R. Epi Info, version 5: a word processing, database, and statistics program for epidemiology on microcomputers. Stone Mountain, Georgia: USD, Inc, 1990. 14. Carlan S, Gore M, VanMeter S, Mastrogiannis D. A longitudinal study evaluating the effect of gestational age on antenatal assessment tests. AM J OBSTETGYNECOL 1992; 166(suppl):412. 15. Moore T, Piacquadio K. A prospective evaluation of fetal movement screening to reduce the incidence of anteparturn fetal death. 1989;160:1075-80. 16. Manning F, Morrison I, Lange I, Harman C. Fetal biophysical profile scoring: selective use of the nonstress test. 1987;156:709-12. 17. Cunningham F, MacDonald E Gant N, Leveno K, Gilstrap L. Williams' obstetrics. 19th ed. Norwalk, Connecticut: Appleton & Lange, 1993:5-6.