AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2 March 1991
FETAL DEATH AFTER NORMAL BIOPHYSICAL PROFILE William J. Watson, M.D., Vern L. Katz, M.D., and Watson A. Bowes, Jr., M.D.
ABSTRACT
The search for the perfect antenatal test has proved an illusory quest in modern obstetrics. The biophysical profile (BPP) is a widely used antepartum test that has an excellent predictive value for fetal well-being. Of special concern to the obstetrician is the rare occurrence of a false-negative test, a fetal death that occurs within 1 week of a normal test, unrelated to a fetal anomaly. We present three cases of fetal death occurring within 3 days of normal BPP tests. CASE REPORTS
Case 1
The patient is a 27-year-old white woman with a twin gestation at 24 weeks' who was admitted to the hospital for evaluation of mild preeclampsia. Her pregnancy had been uncomplicated prior to admission, with concordant growth of both twins. She had no history of hypertension, collagen vascular disease, or renal disease. Admission ultrasound showed concordant twin growth, appropriate for gestational age, with normal amniotic fluid volume in both sacs. Her initial blood pressure was 145/90 torr. She had 2+ pitting edema in the lower extremities, and 1 + protein in the urine. The hematocrit was 33% with 184,000 platelets/mm3. Liver function tests and clotting studies were normal. A 24-hour urine collection revealed 3.6 gm protein. An antinuclear antibody test was negative. A test for the lupus anticoagulant was negative, as was a screening test for gestational diabetes. Because of extreme prematurity, the pa-
tient was managed conservatively and closely observed at bed rest. The following week, repeat 24hour urine collection showed 4.2 gm protein. Fetal surveillance included kick counts and weekly BPP tests. At 26 weeks' gestation, an ultrasound was done to estimate fetal weight; twin A had an estimated weight of 713 gm, that of twin B was 836 gm, a discordance of 15%. No further deterioration of the patient's condition occurred until 27 weeks' gestation, at which time a 24-hour urine showed 7.3 gm of protein. Fetal surveillance was increased to daily BPP testing. Fetal heart rate testing was never reactive, but no decelerations were noted. Doppler studies of the umbilical arteries were normal for this gestational age. Twin A had a systolic to diastolic ratio of 3.3, and twin B had a measured ratio of 3.0. At 282/7 weeks, the patient complained of decreased fetal motion. Ultrasound revealed the death of twin A. The BPP score had been 8 for both infants less than 24 hours before. At delivery, the living twin weighed 930 gm, the other infant weighed 590 gm. There was no evidence of abruption or cord accident. Twin B's neonatal course was complicated by severe respiratory distress syndrome and a grade 4 intracranial hemorrhage. Case 2
The patient is a 22-year-old black woman gravida 2 para 1-0-0-1 at 39 weeks' gestation. She had a prior low transverse cesarean delivery and desired vaginal delivery. She presented to labor and delivery with contractions and was found to be 2 cm dilated and 50% effaced; there was no cervical change over
Maternal Fetal Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Reprint requests: Dr. Watson, Maternal Fetal Medicine, UNC School of Medicine, CB 7570 MacNider, Chapel Hill NC 27514
94
Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
Downloaded by: NYU. Copyrighted material.
Presented are three cases of fetal death, all of which occurred within 3 days of a normal biophysical profile. Despite the excellent record of this antepartum test in predicting fetal well-being, there is a low incidence of false-negative tests; increasing the frequency of testing will not prevent all false-negative tests.
DEATH AFTER NORMAL BIOPHYSICAL PROFILE/Watson, Katz, Bowes
Case 3 The patient is a 27-year-old black woman gravida 3 para 2-0-0-2 with chronic hypertension, renal disease, and a seizure disorder, who was brought to the hospital at 26 weeks' gestation after a grand mal seizure. Her blood pressure was noted to be 200/150 torr. The patient, who had previously been on antihypertensive and antiseizure medications, had not taken her medication for several weeks. Therapy consisted initially of intravenous magnesium sulfate and a hydralazine drip for blood pressure control. Initial laboratory data included a creatinine of 3.3 mg/dl and a blood urea nitrogen of 52 mg/dl. A 24hour urine collection showed 979 mg of protein, and a creatinine clearance of 15 ml/min. (A review of the patient's records revealed a creatinine clearance of 25 ml/min 2 years prior to this time.) Liver enzymes and clotting studies were normal, as was a random serum glucose. On ultrasound study, the fetal biometric parameters were consistent with 26 weeks' gestation, and the estimated weight was 825 gm. After initial stabilization, her blood pressure was controlled with oral hydralazine, atenolol, and clonidine. Following consultation with neurology, it was agreed that her seizure was secondary to uncontrolled chronic hypertension and her seizure disorder, and not secondary to eclampsia. Phenytoin was reinstituted for seizure prophylaxis. The patient was closely observed in the hospital. Her blood pressure was maintained near 140/95 on this regimen. A repeat 24-hour urine collection after 1 week at bed rest in the hospital showed 1160 mg protein. The blood urea nitrogen remained in the range of 38 to 52 mg/ dl, and creatinine values ranged from 2.7 to 3.3. The serum electrolytes were normal. An antinuclear antibody test was negative. Fetal surveillance consisted of daily fetal heart rate monitoring, fetal kick counts, and biweekly BPP examinations. Repeat ultrasound study after 16 days in the hospital showed an increase in estimated fetal weight of 100 gm, with a BPP score of 8. Three days later, the patient reported de-
creased fetal movement, and fetal death was confirmed by ultrasound. The patient subsequently delivered a 950 gm stillborn female. There was no evidence of abruption or cord accident. Autopsy study showed normal anatomy. The placenta weighed 170 gm and was noted to have a 2 cm old infarction. The Kleihauer-Betke test was negative for fetalmaternal hemorrhage. DISCUSSION
The search for the ideal antepartum test is a long, and yet unfinished chapter in modern obstetrics. The nonstress test (NST), contraction stress test (CST), and the BPP are all currently used for antepartum assessment of the fetus. Of special interest to the obstetrician is the specificity of a test used to predict fetal well-being. The ideal antepartum test would have no false-negative results, in which a fetal death occurs after a normal test. Barss et al1 found a fetal death rate of 2.8 per 1000 within 1 week of a normal NST. Smith et al2 noted a similar fetal death rate of 2.6 per 1000 within 7 days of a normal NST and suggested that this rate could be reduced if this test was done in conjunction with a determination of amniotic fluid volume. This modified antepartum test is now widely used in the United States. Freeman and coworkers3 compared the NST to the CST and noted that the CST had a lower false-negative rate; there were only 1.1 fetal deaths per 1000 in 4626 high-risk pregnancies. The BPP was introduced by Manning et al4 in 1980. This test, composed of ultrasound documentation of fetal breathing, fetal movement, amniotic fluid volume, as well as the NST,4 may be superior to the CST in predicting fetal well-being. Clinical experience with the BPP indicates that this test, if normal, is a reliable indicator of fetal well-being. The BPP is often used as a confirmatory test in the case of a nonreactive NST. Thus, the specificity of the BPP is of particular interest. Manning et al3 found a very low incidence of false-negative BPP tests in a population of 19,221 referred high risk pregnancies. There were only 14 fetal deaths in the study population, a rate of 0.7 fetal deaths per 1000 if the test was normal within the prior 7 days. Only 5 of the 14 deaths occurred within 3 days of a normal test, and one of these was a cord prolapse that occurred in a patient at home. Based on these results, Manning et al suggested that twice weekly testing may be indicated in the case of proteinuric hypertension or in an immature fetus with poor growth. Two of our cases fall into this category. Performance of the BPP requires that the examiner be experienced in the use of ultrasound. In the three present cases, one study was performed by a member of the maternal fetal medicine faculty, and the other two were done by chief residents on the obstetric service. All ultrasound examinations are
Downloaded by: NYU. Copyrighted material.
the next 2 hours. The fetal monitor strip showed accelerations, without any periodic decelerations. The patient was seen in clinic the following day, and a BPP was done with a score of 10. Two days later, she presented to the labor suite with contractions; no fetal heart tones were audible. Ultrasound confirmed fetal death. After 12 hours of induced labor, she delivered a stillborn 3370 gm female in a face presentation. A tight nuchal cord required surgical reduction. Autopsy of the infant revealed normal anatomy. The placenta was normal on gross and microscopic examination. Cultures were negative for group B Streptococcus. A screening test for gestational diabetes done at 28 weeks' gestation was negative. A Kleihauer-Betke test showed no evidence of fetal-maternal hemorrhage, and a test for the lupus anticoagulant was negative.
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
96
interesting that in the report by Manning et al,5 5 of the 14 false-negative BPP tests occurred in the clinical setting of intrauterine growth retardation. We report these three cases to enhance the body of information regarding false-negative antenatal testing. There is no current antepartum test that guarantees fetal survival, no matter how often this test is performed. These tests may be less reliable when applied to clinical situations of more severe maternal disease. Perhaps different maternal pathologic conditions will require different types of fetal surveillance. Proteinuric hypertension and growth retardation may be such situations.5 Various gestational ages may also require different testing tools. As more cases of false-negative antenatal tests are reported, we may be able to understand better which tests should be performed for each pregnancy complication. The obstetrician must continue to weigh all factors, including the rare occurrence of a falsenegative BPP, when deciding between delivery or continued expectant management.
REFERENCES
Barss VA, Frigoletto FD, Diamond F: Stillbirth after nonstress testing. Obstet Gynecol 65:541-544, 1985 Smith CV, Nguyen HN, Kovaks B, McCart D, Phelan JP, Paul RH: Fetal death following antepartum fetal heart rate testing: A review of 65 cases. Obstet Gynecol 70:18—20, 1987 Freeman RK, Anderson G, Dorchester W: A prospective multi-institutional study of antepartum fetal heart rate monitoring I. Risk of perinatal mortality and morbidity according to fetal heart rate test results. Am J Obstet Gynecol 143:771-777, 1982 Manning FA, Platt LD, Sipos L: Antepartum fetal evaluation: Development of a fetal biophysical profile. Am J Obstet Gynecol 136:787-795, 1980 Manning FA, Morrison I, Harman CR, Lange IR, Menticoglou S: Fetal assessment based on fetal biophysical profile scoring: Experience in 19,221 referred high risk pregnancies II. An analysis of false-negative fetal deaths. Am J Obstet Gynecol 157:880-884, 1987 Baskett TF, Allen AC, Gray JH, Young DC, Young LM: Fetal biophysical profile and perinatal death. Obstet Gynecol 70:357-360, 1987 Manning FA, Morrison I, Lange IR, Harman CR, Chamberlain PF: Fetal biophysical profile scoring: Selective use of the nonstress test. Am J Obstet Gynecol 156:709-712, 1987
Downloaded by: NYU. Copyrighted material.
recorded and reviewed by a maternal fetal medicine faculty member. Thus, it is unlikely that these BPP tests were falsely interpreted as normal. Baskett et al6 found a false-negative rate of 0.7 per 1000 in applying the BPP to 5034 high-risk pregnancies. Of three deaths after a normal BPP, two were unexplained, and the other was attributed to perinatal asphyxia. The frequency of testing in this study varied from daily to weekly, depending on the clinical situation. The time between the normal BPP and fetal death in these three cases was not stated. Analysis of our three cases reveals a number of important clinical points. The fetal death in case 2 is unexplained, but may have been due to the tight nuchal cord. It is interesting that variable decelerations were not seen on fetal monitoring. The BPP is a test that screens for uteroplacental insufficiency; if this death represented a sudden cord accident, the BPP may not be sufficient to alert the clinician to the problem. If the clinician elects to use the ultrasound for BPP testing without the NST, a cord compression may be missed. Selective use of the BPP without the NST, however, did not decrease predictive test accuracy in a recent study of 2712 high-risk pregnancies.7 The death of one twin in the patient with severe preeclampsia at 28 weeks occurred within 1 day of a normal BPP. Many clinicians would have elected to deliver this patient when criteria for severe preeclampsia were present, despite the extreme fetal immaturity. There is little experience with antepartum testing in the clinical setting of severe preeclampsia, since these patients are usually delivered. It may be that the BPP and other antepartum tests have a poorer predictive value in this setting. The results from antepartum testing in other high-risk situations may not be applicable to the severe preeclamptic patient. Baskett et al6 suggested that the use of Doppler velocimetry might decrease the incidence of falsenegative tests. A normal fetal Doppler study in this patient was observed the week before fetal death occurred. In the case of the mother with severe hypertension and renal disease, the fetus died despite close in-patient observation. Ultrasound after a 2-week period showed only 100 gm of fetal growth. It is
March 1991
FETAL DEATH AFTER NORMAL BIOPHYSICAL PROFILE William J. Watson, M.D., Vern L. Katz, M.D., and Watson A. Bowes, Jr., M.D.
ABSTRACT
The search for the perfect antenatal test has proved an illusory quest in modern obstetrics. The biophysical profile (BPP) is a widely used antepartum test that has an excellent predictive value for fetal well-being. Of special concern to the obstetrician is the rare occurrence of a false-negative test, a fetal death that occurs within 1 week of a normal test, unrelated to a fetal anomaly. We present three cases of fetal death occurring within 3 days of normal BPP tests. CASE REPORTS
Case 1
The patient is a 27-year-old white woman with a twin gestation at 24 weeks' who was admitted to the hospital for evaluation of mild preeclampsia. Her pregnancy had been uncomplicated prior to admission, with concordant growth of both twins. She had no history of hypertension, collagen vascular disease, or renal disease. Admission ultrasound showed concordant twin growth, appropriate for gestational age, with normal amniotic fluid volume in both sacs. Her initial blood pressure was 145/90 torr. She had 2+ pitting edema in the lower extremities, and 1 + protein in the urine. The hematocrit was 33% with 184,000 platelets/mm3. Liver function tests and clotting studies were normal. A 24-hour urine collection revealed 3.6 gm protein. An antinuclear antibody test was negative. A test for the lupus anticoagulant was negative, as was a screening test for gestational diabetes. Because of extreme prematurity, the pa-
tient was managed conservatively and closely observed at bed rest. The following week, repeat 24hour urine collection showed 4.2 gm protein. Fetal surveillance included kick counts and weekly BPP tests. At 26 weeks' gestation, an ultrasound was done to estimate fetal weight; twin A had an estimated weight of 713 gm, that of twin B was 836 gm, a discordance of 15%. No further deterioration of the patient's condition occurred until 27 weeks' gestation, at which time a 24-hour urine showed 7.3 gm of protein. Fetal surveillance was increased to daily BPP testing. Fetal heart rate testing was never reactive, but no decelerations were noted. Doppler studies of the umbilical arteries were normal for this gestational age. Twin A had a systolic to diastolic ratio of 3.3, and twin B had a measured ratio of 3.0. At 282/7 weeks, the patient complained of decreased fetal motion. Ultrasound revealed the death of twin A. The BPP score had been 8 for both infants less than 24 hours before. At delivery, the living twin weighed 930 gm, the other infant weighed 590 gm. There was no evidence of abruption or cord accident. Twin B's neonatal course was complicated by severe respiratory distress syndrome and a grade 4 intracranial hemorrhage. Case 2
The patient is a 22-year-old black woman gravida 2 para 1-0-0-1 at 39 weeks' gestation. She had a prior low transverse cesarean delivery and desired vaginal delivery. She presented to labor and delivery with contractions and was found to be 2 cm dilated and 50% effaced; there was no cervical change over
Maternal Fetal Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Reprint requests: Dr. Watson, Maternal Fetal Medicine, UNC School of Medicine, CB 7570 MacNider, Chapel Hill NC 27514
94
Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
Downloaded by: NYU. Copyrighted material.
Presented are three cases of fetal death, all of which occurred within 3 days of a normal biophysical profile. Despite the excellent record of this antepartum test in predicting fetal well-being, there is a low incidence of false-negative tests; increasing the frequency of testing will not prevent all false-negative tests.
DEATH AFTER NORMAL BIOPHYSICAL PROFILE/Watson, Katz, Bowes
Case 3 The patient is a 27-year-old black woman gravida 3 para 2-0-0-2 with chronic hypertension, renal disease, and a seizure disorder, who was brought to the hospital at 26 weeks' gestation after a grand mal seizure. Her blood pressure was noted to be 200/150 torr. The patient, who had previously been on antihypertensive and antiseizure medications, had not taken her medication for several weeks. Therapy consisted initially of intravenous magnesium sulfate and a hydralazine drip for blood pressure control. Initial laboratory data included a creatinine of 3.3 mg/dl and a blood urea nitrogen of 52 mg/dl. A 24hour urine collection showed 979 mg of protein, and a creatinine clearance of 15 ml/min. (A review of the patient's records revealed a creatinine clearance of 25 ml/min 2 years prior to this time.) Liver enzymes and clotting studies were normal, as was a random serum glucose. On ultrasound study, the fetal biometric parameters were consistent with 26 weeks' gestation, and the estimated weight was 825 gm. After initial stabilization, her blood pressure was controlled with oral hydralazine, atenolol, and clonidine. Following consultation with neurology, it was agreed that her seizure was secondary to uncontrolled chronic hypertension and her seizure disorder, and not secondary to eclampsia. Phenytoin was reinstituted for seizure prophylaxis. The patient was closely observed in the hospital. Her blood pressure was maintained near 140/95 on this regimen. A repeat 24-hour urine collection after 1 week at bed rest in the hospital showed 1160 mg protein. The blood urea nitrogen remained in the range of 38 to 52 mg/ dl, and creatinine values ranged from 2.7 to 3.3. The serum electrolytes were normal. An antinuclear antibody test was negative. Fetal surveillance consisted of daily fetal heart rate monitoring, fetal kick counts, and biweekly BPP examinations. Repeat ultrasound study after 16 days in the hospital showed an increase in estimated fetal weight of 100 gm, with a BPP score of 8. Three days later, the patient reported de-
creased fetal movement, and fetal death was confirmed by ultrasound. The patient subsequently delivered a 950 gm stillborn female. There was no evidence of abruption or cord accident. Autopsy study showed normal anatomy. The placenta weighed 170 gm and was noted to have a 2 cm old infarction. The Kleihauer-Betke test was negative for fetalmaternal hemorrhage. DISCUSSION
The search for the ideal antepartum test is a long, and yet unfinished chapter in modern obstetrics. The nonstress test (NST), contraction stress test (CST), and the BPP are all currently used for antepartum assessment of the fetus. Of special interest to the obstetrician is the specificity of a test used to predict fetal well-being. The ideal antepartum test would have no false-negative results, in which a fetal death occurs after a normal test. Barss et al1 found a fetal death rate of 2.8 per 1000 within 1 week of a normal NST. Smith et al2 noted a similar fetal death rate of 2.6 per 1000 within 7 days of a normal NST and suggested that this rate could be reduced if this test was done in conjunction with a determination of amniotic fluid volume. This modified antepartum test is now widely used in the United States. Freeman and coworkers3 compared the NST to the CST and noted that the CST had a lower false-negative rate; there were only 1.1 fetal deaths per 1000 in 4626 high-risk pregnancies. The BPP was introduced by Manning et al4 in 1980. This test, composed of ultrasound documentation of fetal breathing, fetal movement, amniotic fluid volume, as well as the NST,4 may be superior to the CST in predicting fetal well-being. Clinical experience with the BPP indicates that this test, if normal, is a reliable indicator of fetal well-being. The BPP is often used as a confirmatory test in the case of a nonreactive NST. Thus, the specificity of the BPP is of particular interest. Manning et al3 found a very low incidence of false-negative BPP tests in a population of 19,221 referred high risk pregnancies. There were only 14 fetal deaths in the study population, a rate of 0.7 fetal deaths per 1000 if the test was normal within the prior 7 days. Only 5 of the 14 deaths occurred within 3 days of a normal test, and one of these was a cord prolapse that occurred in a patient at home. Based on these results, Manning et al suggested that twice weekly testing may be indicated in the case of proteinuric hypertension or in an immature fetus with poor growth. Two of our cases fall into this category. Performance of the BPP requires that the examiner be experienced in the use of ultrasound. In the three present cases, one study was performed by a member of the maternal fetal medicine faculty, and the other two were done by chief residents on the obstetric service. All ultrasound examinations are
Downloaded by: NYU. Copyrighted material.
the next 2 hours. The fetal monitor strip showed accelerations, without any periodic decelerations. The patient was seen in clinic the following day, and a BPP was done with a score of 10. Two days later, she presented to the labor suite with contractions; no fetal heart tones were audible. Ultrasound confirmed fetal death. After 12 hours of induced labor, she delivered a stillborn 3370 gm female in a face presentation. A tight nuchal cord required surgical reduction. Autopsy of the infant revealed normal anatomy. The placenta was normal on gross and microscopic examination. Cultures were negative for group B Streptococcus. A screening test for gestational diabetes done at 28 weeks' gestation was negative. A Kleihauer-Betke test showed no evidence of fetal-maternal hemorrhage, and a test for the lupus anticoagulant was negative.
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
96
interesting that in the report by Manning et al,5 5 of the 14 false-negative BPP tests occurred in the clinical setting of intrauterine growth retardation. We report these three cases to enhance the body of information regarding false-negative antenatal testing. There is no current antepartum test that guarantees fetal survival, no matter how often this test is performed. These tests may be less reliable when applied to clinical situations of more severe maternal disease. Perhaps different maternal pathologic conditions will require different types of fetal surveillance. Proteinuric hypertension and growth retardation may be such situations.5 Various gestational ages may also require different testing tools. As more cases of false-negative antenatal tests are reported, we may be able to understand better which tests should be performed for each pregnancy complication. The obstetrician must continue to weigh all factors, including the rare occurrence of a falsenegative BPP, when deciding between delivery or continued expectant management.
REFERENCES
Barss VA, Frigoletto FD, Diamond F: Stillbirth after nonstress testing. Obstet Gynecol 65:541-544, 1985 Smith CV, Nguyen HN, Kovaks B, McCart D, Phelan JP, Paul RH: Fetal death following antepartum fetal heart rate testing: A review of 65 cases. Obstet Gynecol 70:18—20, 1987 Freeman RK, Anderson G, Dorchester W: A prospective multi-institutional study of antepartum fetal heart rate monitoring I. Risk of perinatal mortality and morbidity according to fetal heart rate test results. Am J Obstet Gynecol 143:771-777, 1982 Manning FA, Platt LD, Sipos L: Antepartum fetal evaluation: Development of a fetal biophysical profile. Am J Obstet Gynecol 136:787-795, 1980 Manning FA, Morrison I, Harman CR, Lange IR, Menticoglou S: Fetal assessment based on fetal biophysical profile scoring: Experience in 19,221 referred high risk pregnancies II. An analysis of false-negative fetal deaths. Am J Obstet Gynecol 157:880-884, 1987 Baskett TF, Allen AC, Gray JH, Young DC, Young LM: Fetal biophysical profile and perinatal death. Obstet Gynecol 70:357-360, 1987 Manning FA, Morrison I, Lange IR, Harman CR, Chamberlain PF: Fetal biophysical profile scoring: Selective use of the nonstress test. Am J Obstet Gynecol 156:709-712, 1987
Downloaded by: NYU. Copyrighted material.
recorded and reviewed by a maternal fetal medicine faculty member. Thus, it is unlikely that these BPP tests were falsely interpreted as normal. Baskett et al6 found a false-negative rate of 0.7 per 1000 in applying the BPP to 5034 high-risk pregnancies. Of three deaths after a normal BPP, two were unexplained, and the other was attributed to perinatal asphyxia. The frequency of testing in this study varied from daily to weekly, depending on the clinical situation. The time between the normal BPP and fetal death in these three cases was not stated. Analysis of our three cases reveals a number of important clinical points. The fetal death in case 2 is unexplained, but may have been due to the tight nuchal cord. It is interesting that variable decelerations were not seen on fetal monitoring. The BPP is a test that screens for uteroplacental insufficiency; if this death represented a sudden cord accident, the BPP may not be sufficient to alert the clinician to the problem. If the clinician elects to use the ultrasound for BPP testing without the NST, a cord compression may be missed. Selective use of the BPP without the NST, however, did not decrease predictive test accuracy in a recent study of 2712 high-risk pregnancies.7 The death of one twin in the patient with severe preeclampsia at 28 weeks occurred within 1 day of a normal BPP. Many clinicians would have elected to deliver this patient when criteria for severe preeclampsia were present, despite the extreme fetal immaturity. There is little experience with antepartum testing in the clinical setting of severe preeclampsia, since these patients are usually delivered. It may be that the BPP and other antepartum tests have a poorer predictive value in this setting. The results from antepartum testing in other high-risk situations may not be applicable to the severe preeclamptic patient. Baskett et al6 suggested that the use of Doppler velocimetry might decrease the incidence of falsenegative tests. A normal fetal Doppler study in this patient was observed the week before fetal death occurred. In the case of the mother with severe hypertension and renal disease, the fetus died despite close in-patient observation. Ultrasound after a 2-week period showed only 100 gm of fetal growth. It is
March 1991
