- Email: [email protected]
Evaluation of ultrasound diagnosis of fetal anomalies in women with pregestational diabetes: University of Florida experience
Evaluation of ultrasound diagnosis of fetal anomalies in women with pregestational diabetes: University of Florida experience KevinJ. Gomez, MD, Kaye Dowdy, MSIV, Gail Allen, RN, RDMS, Marcia Tyson-Thomas, RN, RDMS, and Amelia C. Cruz, MD Gainesville, Florida Fetal congenital anomalies have become the major cause of perinatal morbidity and mortality in pregnancies complicated by insulin-dependent diabetes. We evaluated the use of level II ultrasound in predicting congenital anomalies, to determine if the management of pregnant women with insulin-dependent diabetes would be altered by these findings. We examined 43 insulin-dependent diabetic pregnancies. In this group, 10 newborns (23%) were diagnosed either at birth or later to have an abnormality. Of these, four (9%) were diagnosed by ultrasound. Of the seven cases that were undiagnosed, three women did not undergo the level II examination, two women had lesions undetectable by ultrasound, and in two women cardiac abnormalities were missed. Of the four congenital anomalies that were prenatally diagnosed, the findings influenced the management in three cases. The level II ultrasound used as a screening test had a 67% sensitivity and a 100% specificity, with a positive predictive value of 100% and a negative predictive value of 91 %. We conclude that high-resolution ultrasound may be used as a screening tool for congenital anomalies in the insulin-dependent diabetic pregnancy. This study also suggests that the use of fetal echocardiography with evaluation of aortic and pulmonary outflow tracts, the arch of the aorta, and ventricular size may be helpful in the diagnosis of fetal cardiac anomalies in the pregnant woman with insulin-dependent diabetes. (AM J OBSTET GYNECOL 1988;159:584-6.)
Key words: Insulin-dependent diabetes in pregnancy, fetal cardiac anomalies, level II ultrasound
Coincident with the improved management of insulin-dependent diabetic pregnancies, we have seen a decrease in maternal and fetal morbidity and mortality. This can be attributed to tight metabolic control and improved fetal surveillance techniques. In the past, congenital anomalies were responsible for a small percentage of perinatal deaths. 1 However, congenital anomalies have surpassed respiratory distress syndrome as a major cause of perinatal death, accounting for 30% to 50% of perinatal deaths!· 3 With the use of high-resolution ultrasound in the second trimester, one may .better diagnose these congenital anomalies and therefore affect the management of these pregnancies. The purposes of this study were to evaluate the use of level II ultrasound at our institution in the diagnoses of congenital anomalies and to determine how the manFrom the Department of Obstetrics and Gynecology, University of Florida College of Medicine. Supported in part by a grant from the Department of Health and Rehabilitative Services, State of Florida, Children's Medical Services. Presented at the Eighth Annual Meeting of the Society of Perinatal Obstetricians, Las Vegas, Nevada, February 4-6, 1988. Reprint requests: Kevin Gomez, MD, Department of Obstetrics and Gynecology, University of Florida College of Medicine, Box J-294, ]HMHC, Gainesville, FL 32610.
584
agement of these insulin-dependent diabetic pregnancies is altered by these findings.
Material and methods This prospective study was performed from June 1984 to July 1987 at Shands Hospital, University of Florida, Gainesville, Fla. Forty-three women with insulin-dependent diabetes from our clinic population were referred for dating ultrasound at < 14 weeks, level II ultrasound at 14 to 25 weeks, and ultrasound for growth after 25 weeks of gestation. Seventy percent (24 of 34) of the women with insulin-dependent diabetes for whom information on glucose control was available before 20 weeks of gestation were classified as having poor control. Poor control was defined as a hemoglobin Ale level >6.0% by our laboratory or persistently elevated fasting blood sugar levels> 160 mg/dl or 2-hour postprandial blood sugar levels >200 mg/dl. Twentyseven women with insulin-dependent diabetes underwent a level II ultrasound at 14 to 25 weeks of gestation. Sixteen other women with insulin-dependent diabetes did not get level II ultrasound examinations because they failed to show up for scheduled appointments. The ultrasound examinations were performed by three experienced ultrasonographers, one of whom is a peri-
Ultrasound prediction of fetal anomalies in pregestational diabetes
Volume 159 Number 3
Table I. Congenital anomalies Patient No.
2 3 4 5 6 7 8
9
10
Anomalies
Preductal coarctation up aorta, PDA Encephalocele with ventriculomegaly Transposition of great vessels, ASD, VSD, hydrops Small VSD Bilateral talipes equinovarus Cleft upper gum Hemivertebral body Tll without rib Preductal coarctation of aorta, paralysis of right hemidiaphragm Right ventrical hypoplasia, atretic pulmonary trunk, ASD, PDA Phocomelia
585
Table II. Anomalies at birth Initial hemoglobin AI c
NA* 9.0% at 9 wk 8.7% at 17 wk 14.8% at 5 wk 10.9% at 14 wk 7.5% at 13 wk 7.8% at 20 wk 6.0% at 15 wk
9.0% at 7 wk
4.3% at 13 wkt
PDA, Patent ductus arteriosas; ASD, atrial septal defect; VSD, ventricular septal defect.
*All fasting and 2-hour postprandial blood sugar levels were within normal limits. tAll fasting blood sugar levels were > 160 mg/ dl and 2-hour postprandial levels were >260 mg/dl at 7 weeks of gestation.
natologist. Ultrasound examinations were performed with the General Electric model RT 3000. Dating ultrasound consisted of (1) fetal cephalic and body measurements, (2) fetal number, (3) fetal cardiac activity, (4) placental location, and (5) amniotic fluid volume. Level II ultrasound consisted of targeted imaging for fetal anomalies of (1) central nervous system, (2) cardiovascular system, (3) gastrointestinal tract, (4) skeleton, (5) genitourinary tract, and (6) abdominal wall. Cardiac evaluation in level II ultrasound consisted of a four-chamber view with evaluation of cardiac rhythm and calculation of a heart-to-thoracic ratio. The study time for level II ultrasound averaged 30 to 60 minutes. Information on outcome of these insulin-dependent diabetic pregnancies was obtained from neonatal chart review. Outcome results were available for all 43 patients. Results
The characteristics of the study population included 29 multiparous and 15 nulliparous patients with a mean age of 27.6 years (range 18 to 42 years). The distribution of insulin-dependent diabetes according to White's classification revealed that 20 subjects were class B, 6 were class C, 6 were class D, 1 was class R, 7 were class F, 2 were class RF, and 1 was class HT. The incidence of infants with congenital anomalies by White's
Positive
Ultrasound diagnosis Positive 4 Negative 2 Total 6 Sensitivity = 67% Specificity = 100% Positive predictive value = 100% Negative predictive value = 91%
Negative
0
21 21
Total
4
23
classification showed that there were 4 (20%) in class B, 1 (17%) in class C, 1 (17%) in class D, 3 (42%) in class F, 0 in class RF, 1 ( 100%) in class HT, and 0 in class R. A total of 10 babies had congenital anomalies identified either at birth or later in this group of 43 insulindependent pregnant women, giving an overall incidence of 23%. Of these, four were diagnosed by ultrasound. Anomalies included (1) phocomelia, (2) encephalocele with ventriculomegaly, (3) bilateral talipes equinovarus, and (4) increased heart-to-thoracic ratio, 2-to-1 atrioventricular dissociation, fetal hydrops, and bradycardia. Of the six cases not diagnosed by ultrasound, there were three patients who did complete the protocol (no Level II examination). In this group the anomalies were as follows: (1) ventricular septal defect, (2) right ventricular hypoplasia, atrial septal defect, patent ductus arteriosus, and pulmonary atresia, and (3) absence of hemivertebral body and left rib at T11. There was one case of a cleft upper gum that was not diagnosable by ultrasound. Two cases were missed by ultrasound; both of these were coarctation of the aorta. The highest incidences of congenital anomalies were found in the cardiovascular system (50%) and the skeletal system (30% ). The two congenital anomalies missed prenatally by ultrasound, as noted earlier, were in the cardiovascular system. These two newborns were found to have preductal coarctation of the aorta. One woman had undergone ultrasound examination at 8, 12, and 32 weeks of gestation, whereas the other had the evaluation done at 8, 20, and 22 weeks. Table I lists all the congenital anomalies that were diagnosed prenatally and after delivery. When the level II screening ultrasound at 14 to 25 weeks of gestation was used, four of six detectable anomalies were diagnosed, with a 67% sensitivity (Table II). All the detected anomalies were confirmed (100% specificity). The positive predictive value is therefore 100% and the negative predictive value is 91%. The management of these patients was influenced in three cases due to the findings on ultrasound as follows.
586 Gomez et al.
Case reports Case 1. A 29-year-old woman (para 0-0-1-0) with HT class insulin-dependent diabetes had a level II ultrasound performed at 22 weeks of gestation that revealed encephalocele. The patient was counselled as to the findings and decided to continue with the pregnancy. On follow-up examinations at 27 and 32 weeks of gestation, hydrocephalus was diagnosed and noted to be increasing, along with preeclampsia; it was decided to perform a primary cesarean section at 32 weeks of gestation. The infant was confirmed to have encephalocele with hydrocephalus at delivery. A ventriculoperitoneal shunt was placed. This child is presently 3 years old and has developmental and physical delays. Case 2. A 23-year-old woman (para 0-1-l-l) with class D diabetes had an initial dating ultrasound that was normal, but a level II ultrasound at 22 weeks of gestation revealed an increased heart-to-thoracic ratio of 72%, with 2: I atrioventricular dissociation, thick ventricular wall, and fetal bradycardia. Follow-up evaluation at 32 weeks of gestation was significant for fetal hydrops, polyhydramnios, and fetal bradycardia. This patient had a cesarean section performed, and a cardiac pacemaker was placed in the neonate soon after birth. Further cardiac evaluation of the neonate revealed transposition of the great vessels, atrial septal defect, ventricular septal defect, and bilateral bilobed lungs. This infant died at 29 days of life. Case 3. A 24-year-old woman (para 0-0-0-0) with class C diabetes had a dating ultrasound at 9 weeks of gestation. A level II ultrasound was done at 22o/7 weeks of gestation, at which time the diagnosis of phocomelia was made. The patient was made aware of the findings, and she decided to terminate the pregnancy via prostaglandin induction. The prenatal diagnosis was confirmed at the time of delivery. Comment Several studies have shown a significant increase in the risk of congenital anomalies in infants of mothers with diabetes!· 5 In our study population of insulindependent pregnancies, the incidence of congenital anomalies was 23%. This incidence is somewhat higher when compared with the 17% incidence noted in the prospective study done by the Collaborative Perinatal Project. 6 Our study revealed the expected range of congenital anomalies, from the rare phocomelic diabetic embryopathy to the most common cardiac malformations. The use of high-resolution ultrasound played a valuable role not only in diagnosis, but also in further management of these pregnancies.
September 1988 Am J Obstet Gynecol
A review of our data on the cardiac malformations showed an incidence of ll% as compared with 2.5% in the Collaborative Perinatal Project. 6 Ventricular septal defect and coarctation of the aorta comprised >50% of the cardiac malformations, as in other reported series.5· 7 The two false negative results were in detection of cardiac malformation. In this high-risk, insulindependent diabetic pregnancy group, the ultrasound four-chamber view was unable to prenatally diagnose two cases of coarctation of the aorta. Cope! et a!. 8 reported the use of the four-chamber view in screening for congenital heart disease. In their series they also reported one case of coarctation of the aorta with a normal four-chamber view. Although Copel et a!. reported a sensitivity of 92% and a specificity of 99.7% for prenatal evaluation of the fetus at risk for congenital heart disease, this may not hold true for the selected group of insulin-dependent diabetic pregnancies with a higher incidence of cardiac anomalies." These results would suggest that the use of fetal echocardiography with evaluation of the aortic and pulmonary outflow tracts, the arch of the aorta, and ventricular size could increase the sensitivity of the screening protocol. These added evaluations are now part of our level II cardiac examination, performed with the collaboration of a pediatric cardiologist at our institution. REFERENCES l. Driscoll SG, Benirschke K, Curtis GW. Neonatal deaths
2. 3. 4. 5. 6. 7. 8.
among infants of diabetic mothers. Am 1 Dis Child 1961 ; 100:8I8. Coustan DR. Recent advances in the management of diabetic pregnant women. C1in Perinato!I980;7:299. Soler NW, Soler SM, Malins 1M. Neonatal morbidity among infants of diabetic mothers. Diabetes Care I978; I :340. Molsted-Pedersen L, Tygstrup Ll, Pedersen J. Congenital malformations in newborn infants of diabetic women. Lancet I964;I:1I24-6. Mitchell SC, Sellmann AH, Westphal MC, Park]. Etiologic correlates in a study of congenital heart disease in 56,I09 births. Am1 Cardia! I971;28:653-7. Chung CS, Myrianthopoulos NC. Factors affecting risks of congenital malformations. Birth Defects I975; II ( 10): 1-28. Rowland TW, Hubbell 1P 1r, Nadas AS. Congenital heart disease in infants of diabetic mothers. 1 Pediatr 1973; 83:815-20. Cope! 1A, Pilu G, Green 1. eta!. Fetal echocardiographic screening for congenital heart disease: The importance of the four-chamber view. AM 1 0BSTET GYNECOL 1987; 157:648-55.
Key words: Insulin-dependent diabetes in pregnancy, fetal cardiac anomalies, level II ultrasound
Coincident with the improved management of insulin-dependent diabetic pregnancies, we have seen a decrease in maternal and fetal morbidity and mortality. This can be attributed to tight metabolic control and improved fetal surveillance techniques. In the past, congenital anomalies were responsible for a small percentage of perinatal deaths. 1 However, congenital anomalies have surpassed respiratory distress syndrome as a major cause of perinatal death, accounting for 30% to 50% of perinatal deaths!· 3 With the use of high-resolution ultrasound in the second trimester, one may .better diagnose these congenital anomalies and therefore affect the management of these pregnancies. The purposes of this study were to evaluate the use of level II ultrasound at our institution in the diagnoses of congenital anomalies and to determine how the manFrom the Department of Obstetrics and Gynecology, University of Florida College of Medicine. Supported in part by a grant from the Department of Health and Rehabilitative Services, State of Florida, Children's Medical Services. Presented at the Eighth Annual Meeting of the Society of Perinatal Obstetricians, Las Vegas, Nevada, February 4-6, 1988. Reprint requests: Kevin Gomez, MD, Department of Obstetrics and Gynecology, University of Florida College of Medicine, Box J-294, ]HMHC, Gainesville, FL 32610.
584
agement of these insulin-dependent diabetic pregnancies is altered by these findings.
Material and methods This prospective study was performed from June 1984 to July 1987 at Shands Hospital, University of Florida, Gainesville, Fla. Forty-three women with insulin-dependent diabetes from our clinic population were referred for dating ultrasound at < 14 weeks, level II ultrasound at 14 to 25 weeks, and ultrasound for growth after 25 weeks of gestation. Seventy percent (24 of 34) of the women with insulin-dependent diabetes for whom information on glucose control was available before 20 weeks of gestation were classified as having poor control. Poor control was defined as a hemoglobin Ale level >6.0% by our laboratory or persistently elevated fasting blood sugar levels> 160 mg/dl or 2-hour postprandial blood sugar levels >200 mg/dl. Twentyseven women with insulin-dependent diabetes underwent a level II ultrasound at 14 to 25 weeks of gestation. Sixteen other women with insulin-dependent diabetes did not get level II ultrasound examinations because they failed to show up for scheduled appointments. The ultrasound examinations were performed by three experienced ultrasonographers, one of whom is a peri-
Ultrasound prediction of fetal anomalies in pregestational diabetes
Volume 159 Number 3
Table I. Congenital anomalies Patient No.
2 3 4 5 6 7 8
9
10
Anomalies
Preductal coarctation up aorta, PDA Encephalocele with ventriculomegaly Transposition of great vessels, ASD, VSD, hydrops Small VSD Bilateral talipes equinovarus Cleft upper gum Hemivertebral body Tll without rib Preductal coarctation of aorta, paralysis of right hemidiaphragm Right ventrical hypoplasia, atretic pulmonary trunk, ASD, PDA Phocomelia
585
Table II. Anomalies at birth Initial hemoglobin AI c
NA* 9.0% at 9 wk 8.7% at 17 wk 14.8% at 5 wk 10.9% at 14 wk 7.5% at 13 wk 7.8% at 20 wk 6.0% at 15 wk
9.0% at 7 wk
4.3% at 13 wkt
PDA, Patent ductus arteriosas; ASD, atrial septal defect; VSD, ventricular septal defect.
*All fasting and 2-hour postprandial blood sugar levels were within normal limits. tAll fasting blood sugar levels were > 160 mg/ dl and 2-hour postprandial levels were >260 mg/dl at 7 weeks of gestation.
natologist. Ultrasound examinations were performed with the General Electric model RT 3000. Dating ultrasound consisted of (1) fetal cephalic and body measurements, (2) fetal number, (3) fetal cardiac activity, (4) placental location, and (5) amniotic fluid volume. Level II ultrasound consisted of targeted imaging for fetal anomalies of (1) central nervous system, (2) cardiovascular system, (3) gastrointestinal tract, (4) skeleton, (5) genitourinary tract, and (6) abdominal wall. Cardiac evaluation in level II ultrasound consisted of a four-chamber view with evaluation of cardiac rhythm and calculation of a heart-to-thoracic ratio. The study time for level II ultrasound averaged 30 to 60 minutes. Information on outcome of these insulin-dependent diabetic pregnancies was obtained from neonatal chart review. Outcome results were available for all 43 patients. Results
The characteristics of the study population included 29 multiparous and 15 nulliparous patients with a mean age of 27.6 years (range 18 to 42 years). The distribution of insulin-dependent diabetes according to White's classification revealed that 20 subjects were class B, 6 were class C, 6 were class D, 1 was class R, 7 were class F, 2 were class RF, and 1 was class HT. The incidence of infants with congenital anomalies by White's
Positive
Ultrasound diagnosis Positive 4 Negative 2 Total 6 Sensitivity = 67% Specificity = 100% Positive predictive value = 100% Negative predictive value = 91%
Negative
0
21 21
Total
4
23
classification showed that there were 4 (20%) in class B, 1 (17%) in class C, 1 (17%) in class D, 3 (42%) in class F, 0 in class RF, 1 ( 100%) in class HT, and 0 in class R. A total of 10 babies had congenital anomalies identified either at birth or later in this group of 43 insulindependent pregnant women, giving an overall incidence of 23%. Of these, four were diagnosed by ultrasound. Anomalies included (1) phocomelia, (2) encephalocele with ventriculomegaly, (3) bilateral talipes equinovarus, and (4) increased heart-to-thoracic ratio, 2-to-1 atrioventricular dissociation, fetal hydrops, and bradycardia. Of the six cases not diagnosed by ultrasound, there were three patients who did complete the protocol (no Level II examination). In this group the anomalies were as follows: (1) ventricular septal defect, (2) right ventricular hypoplasia, atrial septal defect, patent ductus arteriosus, and pulmonary atresia, and (3) absence of hemivertebral body and left rib at T11. There was one case of a cleft upper gum that was not diagnosable by ultrasound. Two cases were missed by ultrasound; both of these were coarctation of the aorta. The highest incidences of congenital anomalies were found in the cardiovascular system (50%) and the skeletal system (30% ). The two congenital anomalies missed prenatally by ultrasound, as noted earlier, were in the cardiovascular system. These two newborns were found to have preductal coarctation of the aorta. One woman had undergone ultrasound examination at 8, 12, and 32 weeks of gestation, whereas the other had the evaluation done at 8, 20, and 22 weeks. Table I lists all the congenital anomalies that were diagnosed prenatally and after delivery. When the level II screening ultrasound at 14 to 25 weeks of gestation was used, four of six detectable anomalies were diagnosed, with a 67% sensitivity (Table II). All the detected anomalies were confirmed (100% specificity). The positive predictive value is therefore 100% and the negative predictive value is 91%. The management of these patients was influenced in three cases due to the findings on ultrasound as follows.
586 Gomez et al.
Case reports Case 1. A 29-year-old woman (para 0-0-1-0) with HT class insulin-dependent diabetes had a level II ultrasound performed at 22 weeks of gestation that revealed encephalocele. The patient was counselled as to the findings and decided to continue with the pregnancy. On follow-up examinations at 27 and 32 weeks of gestation, hydrocephalus was diagnosed and noted to be increasing, along with preeclampsia; it was decided to perform a primary cesarean section at 32 weeks of gestation. The infant was confirmed to have encephalocele with hydrocephalus at delivery. A ventriculoperitoneal shunt was placed. This child is presently 3 years old and has developmental and physical delays. Case 2. A 23-year-old woman (para 0-1-l-l) with class D diabetes had an initial dating ultrasound that was normal, but a level II ultrasound at 22 weeks of gestation revealed an increased heart-to-thoracic ratio of 72%, with 2: I atrioventricular dissociation, thick ventricular wall, and fetal bradycardia. Follow-up evaluation at 32 weeks of gestation was significant for fetal hydrops, polyhydramnios, and fetal bradycardia. This patient had a cesarean section performed, and a cardiac pacemaker was placed in the neonate soon after birth. Further cardiac evaluation of the neonate revealed transposition of the great vessels, atrial septal defect, ventricular septal defect, and bilateral bilobed lungs. This infant died at 29 days of life. Case 3. A 24-year-old woman (para 0-0-0-0) with class C diabetes had a dating ultrasound at 9 weeks of gestation. A level II ultrasound was done at 22o/7 weeks of gestation, at which time the diagnosis of phocomelia was made. The patient was made aware of the findings, and she decided to terminate the pregnancy via prostaglandin induction. The prenatal diagnosis was confirmed at the time of delivery. Comment Several studies have shown a significant increase in the risk of congenital anomalies in infants of mothers with diabetes!· 5 In our study population of insulindependent pregnancies, the incidence of congenital anomalies was 23%. This incidence is somewhat higher when compared with the 17% incidence noted in the prospective study done by the Collaborative Perinatal Project. 6 Our study revealed the expected range of congenital anomalies, from the rare phocomelic diabetic embryopathy to the most common cardiac malformations. The use of high-resolution ultrasound played a valuable role not only in diagnosis, but also in further management of these pregnancies.
September 1988 Am J Obstet Gynecol
A review of our data on the cardiac malformations showed an incidence of ll% as compared with 2.5% in the Collaborative Perinatal Project. 6 Ventricular septal defect and coarctation of the aorta comprised >50% of the cardiac malformations, as in other reported series.5· 7 The two false negative results were in detection of cardiac malformation. In this high-risk, insulindependent diabetic pregnancy group, the ultrasound four-chamber view was unable to prenatally diagnose two cases of coarctation of the aorta. Cope! et a!. 8 reported the use of the four-chamber view in screening for congenital heart disease. In their series they also reported one case of coarctation of the aorta with a normal four-chamber view. Although Copel et a!. reported a sensitivity of 92% and a specificity of 99.7% for prenatal evaluation of the fetus at risk for congenital heart disease, this may not hold true for the selected group of insulin-dependent diabetic pregnancies with a higher incidence of cardiac anomalies." These results would suggest that the use of fetal echocardiography with evaluation of the aortic and pulmonary outflow tracts, the arch of the aorta, and ventricular size could increase the sensitivity of the screening protocol. These added evaluations are now part of our level II cardiac examination, performed with the collaboration of a pediatric cardiologist at our institution. REFERENCES l. Driscoll SG, Benirschke K, Curtis GW. Neonatal deaths
2. 3. 4. 5. 6. 7. 8.
among infants of diabetic mothers. Am 1 Dis Child 1961 ; 100:8I8. Coustan DR. Recent advances in the management of diabetic pregnant women. C1in Perinato!I980;7:299. Soler NW, Soler SM, Malins 1M. Neonatal morbidity among infants of diabetic mothers. Diabetes Care I978; I :340. Molsted-Pedersen L, Tygstrup Ll, Pedersen J. Congenital malformations in newborn infants of diabetic women. Lancet I964;I:1I24-6. Mitchell SC, Sellmann AH, Westphal MC, Park]. Etiologic correlates in a study of congenital heart disease in 56,I09 births. Am1 Cardia! I971;28:653-7. Chung CS, Myrianthopoulos NC. Factors affecting risks of congenital malformations. Birth Defects I975; II ( 10): 1-28. Rowland TW, Hubbell 1P 1r, Nadas AS. Congenital heart disease in infants of diabetic mothers. 1 Pediatr 1973; 83:815-20. Cope! 1A, Pilu G, Green 1. eta!. Fetal echocardiographic screening for congenital heart disease: The importance of the four-chamber view. AM 1 0BSTET GYNECOL 1987; 157:648-55.