- Email: [email protected]
Iliac wing angle as a marker for trisomy 21 in the second trimester
• Iliac Wing Angle as a Marker for Trisomy 21 In the Second Trimester
MICHAEL D. BORK, DO, JAMES F. X. EGAN, MD, WILLIAM CUSICK, MD, ADAM F. BORGIDA, MD, WINSTON A. CAMPBELL, MD, AND JOHN F. RODIS, MD Objective: To evaluate the fetal iliac wing angle in detecting trisomy 21 in the second trimester of pregnancy. Methods: Using an axial view of the fetal pelvis, the angle between the right and left iliac wings (iliac wing angle) was measured ultrasonographically at the time of the second• trimester ultrasound or genetic amniocentesis in 377 single• ton fetuses. Trisomy 21 was diagnosed by karyotype results from the amniocentesis or newborn examination with karyo• type if trisomy 21 was suspected based on phenotypic features. Sensitivity, specificity, and positive (PPV) and negative predictive values (NPV) were calculated using multiple cutoff points. A receiver operating characteristic (ROC) curve was used to identify the optimum iliac wing angle. Descriptive statistics and Student t test were utilized for analyses with a P of less than .05 considered significant. Results: The average gestational age was 18.8 weeks (range 13-32). Karyotypes were available in 128 fetuses. The overall prevalence of trisomy 21 was 11 of 377 (2.9%). The mean (± standard deviation) iliac wing angle in the normal fetuses was 68.2 degrees (±15.4 degrees) and 98.5 degrees (±11.3 degrees) in fetuses with trisomy 21 (P < .001). Using an ROC-derived absolute cutoff of 90 degrees, an abnormal iliac wing angle had sensitivity of 90.9% (ten of 11), specificity of 94.50/. (346 of 366), NPV of 99.7% (346 of 347), and PPV of 33.3% (ten of 30) to detect trisomy 21. Conclusion: Fetuses with trisomy 21 have greater iliac wing angles than do normal fetuses. Using an ROC-derived absolute cutoff of 90 degrees, we could detect 90.9% of fetuses with trisomy 21 with a PPV of 33% in our high-risk population. These findings suggest that iliac wing angle is a useful marker in antenatal screening for trisomy 21. (Obstet GynecoI1997i89:734-7. © 1997 by The American College of Obstetricians and Gynecologists.)
Sonographic markers associated with fetal karyotype abnormalities were limited initially to the detection of major structural malformations. Over the last decade, From the Departments of Obstetrics and Gynecology, Divisions of Maternal-Fetal Medicine, Oakwood Hospital and Medical Center, Dear• born, Michigan, and the University of Connecticut Health Center, Farmington, Connecticut.
734 0029-7844/97/$17.00
PII 50029-7844(97)00089-6
however, more subtle ultrasound markers have been identified in the second trimester of pregnancy to identify a fetus at increased risk for a karyotype abnor• mality. These subtle ultrasound markers include con• genital heart defects, increased nuchal skin thickness, abnormally short or low-set fetal ears, shortened prox• imal long bones (ie, femur and humerus), increased brightness or echogenicity of the fetal bowel, a two• vessel umbilical cord, clubbed feet, pyelectasis, hypo• plasia of the mid-phalanx of the fifth finger, or an increased space between the first and second toes. 1- 3 Trisomy 21 is the most common karyotype abnormal• ity found in newborns, approximately 1:800 in the general population. 4 Up to 80% of newborns with trisomy 21 have abnormalities of the pelvic bones. These include small acetabular angles, elongated taper• ing ischia, and bilateral widening and flaring of the iliac wingS. 5,6 Although many of these pelvic bone abnor• malities have been documented in the neonatal period, their presence and diagnostic utility in the antenatal period has not yet been investigated. Our objective was to evaluate the efficacy of the ultrasonic measurement of fetal iliac wing angle as a marker for trisomy 21 in the second trimester of pregnancy.
Materials and Methods We identified women with singleton gestations who were referred for a targeted ultrasound or genetic amniocentesis from January 1, 1995, to June 30, 1995. Fetuses with previously determined karyotype abnor• malities were excluded. Collected data included mater• nal age, referral indications, gestational age at time of ultrasound, karyotype results, and iliac wing angle measurement. Using an axial (transverse) view of the fetal pelvis, including the sacral spine, the angle between the right and left iliac wings was measured on-screen (Figure 1), and designated the "iliac wing angle." To ensure a true
Obstetrics & Gynecology
Table 1. Indications for Ultrasound Referral Indication
n
(%)
Advanced maternal age Abnormal triple screen Trisomy 21 risk> 1:270 MSAFP > 2.0 MoM Uncertain dates Family history of congenital anomaly Medical complication of pregnancy Medication/teratogen exposure Other
159
(42.2)
44 27
01.7)
59 29 28 19 12
05.6)
m
Th~
(7.2) (7.7)
(7.4)
(5.0) (3.2) ~
MSAFP = maternal serum alpha-fetoprotein; MoM = multiples of the median.
Figure 1. Iliac wing angle determination from axial section through fetal pelvis.
axial section perpendicular to the sacral spine, the image was taken at a level where both iliac bones were at their longest dimension and of approximately equal length. This level generally includes the dome of the fetal bladder, and is always below the umbilical cord insertion. All angle measurements were performed by the same examiner (MOB) using a previously described on-screen method? All measurements were taken and calculated before knowledge of the outcome. The diagnosis of trisomy 21 was made by karyotype results from amniocentesis or by newborn examination for phenotypic evidence of Oown syndrome with karyotype confirmation. The sensitivity, specificity, positive (PPV), and negative predictive values (NPV) for the iliac wing angle measurement were calculated using multiple cutoff points, and a receiver operating characteristic (ROC) curve was generated to identify the optimum iliac wing angle measurement. Interobserver variation was evaluated in 20 consecutive normal fe• tuses by two observers who were blinded to each other's results. Each observer independently obtained an axial image of the fetal pelvis and measured the pelvic angle. The absolute values of pair differences [mean :!:: standard deviation (SO)} were calculated. Our statistical analysis included descriptive statistics, Stu• dent t test for continuous variables, and for categoric data, with P less than .05 deemed significant.
cases. A diagnosis of trisomy 21 was made in 11 (2.9%) of the 377 fetuses. The mean (±SO) iliac wing angle measurement in the normal fetuses was 68.2 degrees (±15.4 degrees) compared with 98.5 degrees (±11.3 degrees) in fetuses with trisomy 21 (P < .00l). The normal iliac wing angle measurement was independent of gestational age (r = 0.04). Figure 2 shows the angle measurement in a normal fetus. Interobserver variation was evaluated in 20 consecutive normal fetuses with a mean (±SD) variation of 4.4 degrees (±2.8 degrees) and absolute differences between matched pairs ranged between 0 degrees and 11 degrees. Using an ROC-derived optimum cutoff point of 90 degrees (Figure 3), ten of the 11 fetuses with trisomy 21 could be identified. The iliac wing angle measurement of 90 degrees or greater had a sensitivity of 90.9% (ten of 11) and a specificity of 94.5% (346 of 366). The PPV
x:
Results Our study population of 377 women had a mean (± SO; range) maternal age of 31.8 (±6.2; 15-44) years. The mean (±SD; range) gestational age of the fetuses was 18.8 (:!::2.4; 13-32) weeks. The referral indications are listed in Table 1. Amniocentesis was performed in 128
VOL. 89, NO.5, PART 1, MAY 1997
Figure 2. Axial section through fetal pelvis demonstrating measure• ment of iliac wing angle (IL = iliac wing; SP = spine).
Bork et al
Iliac Wing Angle
735
1.0 0.9 0.8 0.7
= = >-
~
til
c:
Ql
en
r::-
80'
73'
II'
116 ' 1111'
0.6 0.5 0.4 0.3 0.2 0.1
0.0 ......"""T'"-.-..-.."""T"".......,.......-.-..-,-"'T"""....,.---r-".-l 0.00.1 0.20.30.40.50.60.70.80.91.0
1-speclflclty Figure 3. Receiver operating characteristic curve for fetal iliac wing angle.
was 33.3% (ten of 30), and the NPV was 99.7% (346 of 347) in our population.
Discussion Caffey and Ross 5 were the first to diagnose changes in the pelvic bones of infants with Down syndrome. These findings have been confirmed by several other investi• gators. 8 - 10 The classic features in these reports are bilateral flattening of the lower edges of the ilia, elon• gated tapering ischia, small acetabular angles, and bilateral widening and flaring of the iliac wings. Our findings indicate that some of the characteristic skeletal abnormalities of the pelvis seen in newborns with trisomy 21 are present in the second trimester of pregnancy. We found that the iliac wing angle, which represents the "flaring of the iliac wings" described by Caffey and ROSS,6 is significantly increased in trisomy 21 fetuses when compared with normal ones. Little has been reported previously regarding prena• tal ultrasound evaluation of the fetal ilia. Two re• ports ll •12 of ultrasonographic measurement of fetal iliac bone length in normal fetuses have shown a linear relationship between fetal iliac bone length and gesta• tional age. Abuhamed et al13 have evaluated the utility of ultrasonographic iliac bone length in the identification of fetuses with Down syndrome in the second trimester of pregnancy. They have shown that iliac length in Down syndrome fetuses (longest bony dimension) is significantly longer than that of controls (P < .00l). In their study, 40% of trisomy 21 fetuses were identified by a measured-ta-expected iliac length ratio of at least 1.21. These results differ from our own study which mea-
736 Bark et
at Iliac Wing Angle
sures the angle described by the intersection of lines drawn through an axial cut of two iliac bones. The iliac bones are easily visualized, but technical difficulty in obtaining the proper image plane may occur when either the iliac wings are not of equal size (generally representing lateral obliquity with wings imaged at different heights) or if the intersection with the fetal anterior abdominal wall is below the dome of the bladder or at or above the umbilical cord insertion (suggesting anterior-posterior obliquity). To ensure uni• formity of measurements, we obtained axial sections through the fetal pelvis at the level of the upper sacral spine. This is approximately the same plane Abuhamed et al 13 used to measure the maximum length of the iliac wing. Although the angles in our study were electron• ically measured by calipers, the ROC-derived absolute cutoff of 90 degrees lends itself to visual discrimination of normal from abnormal. The prevalence of trisomy 21 in our study population of 2.9% (11 of 377) reflects our high-risk referral popu• lation, and is a limitation of our study. In a lower-risk population, given a lower prevalence, the false-positive rate would be higher. We extrapolated our data to a lower-risk population using Bayes theorem with the a priori risks derived from the patient's age or her triple screen. 3 For example, an abnormal iliac wing angle measurement increased the risk for trisomy 21 for a pregnant woman 35 years of age from 1 in 270 to 1 in 24. Conversely, until the a priori triple screen risk is less than 1 in 3500, an abnormal iliac wing angle at the time of a second trimester ultrasound will indicate a risk of trisomy 21 that is greater than 1 in 270. Further pra• spective studies in low-risk populations will be neces• sary to evaluate fully the utility of the iliac wing angle as a screening marker for trisomy 21. In our study, the iliac wing angle was 90.9% sensitive in detecting trisomy 21. Compared to the more com• monly recognized second-trimester ultrasonographic markers of aneuploidy, our findings suggest that iliac wing angle measurement will be among the most sensitive. Vintzileos and Egan3 reviewed the English language literature over a to-year period and reported the following mean (range) sensitivities: short femur 31% (13-70%), short humerus 33% (24-64%), and nu• chal fold thickening 34.1% (9-75%). In two of our 11 fetuses with trisomy 21, an increased iliac wing angle was the only ultrasound marker of aneuploidy. A third fetus had an isolated echogenic intracardiac focus of the left ventricle along with the increased iliac wing angle. The echogenic focus has also been described as a possible ultrasound marker for trisomy 21. 14 The re• maining eight fetuses had at least one of the more commonly recognized second-trimester ultrasono• graphic markers of aneuploidy. A priori knowledge by
Obstetrics & Gynecology
the ultrasonographer of these subtle anatomic changes suggestive of trisomy 21 could have introduced bias into the pelvic angle measurement; this may be a limitation of the study. Larger prospective studies in low-risk populations are required to assess accurately the utility of this marker.
References 1. Benacerraf BR, Neuberg 0, Bromley B, Frigoletto FD. Sonographic scoring index for prenatal detection of chromosomal abnormali• ties. J Ultrasound Med 1992;11:449-58. 2. Nyberg DA, Resta RG, Luthy DA, Hickok DE, Mahony B5, Hirsch JH. Prenatal sonographic findings of Down syndrome: Review of 94 cases. Obstet Gynecol 1990;76:370-7. 3. Vintzileos AM, Egan JFX. Adjusting the risk for trisomy 21 on the basis of second-trimester ultrasonography. Am J Obstet Gynecol 1995;172:837-44. 4. Thompson M, McInnes RR, Willard HF. Clinical cytogenetics: General principles and autosomal abnormalities. In: Thompson JS, Thompson MW, eds. Genetics in medicine. 5th ed. Philadelphia: WB Saunders, 1991:201-29. 5. Caffey J, Ross S. Mongolism (mongoloid deficiency) during early infancy: Some newly recognized diagnostic changes in the pelvic bones. Pediatrics 1956;17:642-51. 6. Caffey J, Ross S. Pelvic bones in infantile mongoloidism: Roent• genographic features. Am J RoentgenoI1958;80:458-67. 7. Bork MD, Egan JFX, Diana OJ, Scorza WE, Fabbri EL, Feeney LD, et al. A new method for on-screen ultrasound determination of fetal cardiac axis. Am J Obstet Gynecoll995;173:1192-5.
VOL. 89, NO.5, PART 1, MAY 1997
8. Kaufmann HJ, Taillard WF. Pelvic abnormalities in mongols. BMJ 1961;1:948-9. 9. Hall B. Mongolism in newborns: A clinical and cytogenic study. Acta Paediatr Scand Suppl 1964;154:41-8. 10. Mortensson W, Hall B. Abnormal pelvis in newborn infants with Down's syndrome. Acta RadioI1972;12:847-55. 11. Apuzzio n, Adhate A, Ganesh V, Leo MY, Holland BK. Prenatal ultrasonographic fetal iliac bone measurement: Correlation with gestational age. J Reprod Med 1992;37:348-50. 12. Jaffe R, Santolaya J, Warsof SL. Sonographic measurement of the fetal iliac bone: Growth and relation to femur length in assessment of gestational age. Am J Perinatol 1993;10:105-8. 13. Abuhamed AZ, Kolm P, Marl G, Slotnick RN, Evans AT. Ultra• sonographic fetal iliac length measurement in the screening for Down syndrome. Am J Obstet Gynecol 1994;171:1063-7. 14. Bromley B, Lieberman E, Laboda L, Benacerraf BR. Echogenic intracardiac focus: A sonographic sign for fetal Down syndrome. Obstet Gynecoll995;86:998-1001.
Reprints are not available.
Received November 7, 1996. Received in revised form January 10, 1997. Accepted January 28, 1997.
Copyright © 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Bork et al Iliac Wing Angle
737
MICHAEL D. BORK, DO, JAMES F. X. EGAN, MD, WILLIAM CUSICK, MD, ADAM F. BORGIDA, MD, WINSTON A. CAMPBELL, MD, AND JOHN F. RODIS, MD Objective: To evaluate the fetal iliac wing angle in detecting trisomy 21 in the second trimester of pregnancy. Methods: Using an axial view of the fetal pelvis, the angle between the right and left iliac wings (iliac wing angle) was measured ultrasonographically at the time of the second• trimester ultrasound or genetic amniocentesis in 377 single• ton fetuses. Trisomy 21 was diagnosed by karyotype results from the amniocentesis or newborn examination with karyo• type if trisomy 21 was suspected based on phenotypic features. Sensitivity, specificity, and positive (PPV) and negative predictive values (NPV) were calculated using multiple cutoff points. A receiver operating characteristic (ROC) curve was used to identify the optimum iliac wing angle. Descriptive statistics and Student t test were utilized for analyses with a P of less than .05 considered significant. Results: The average gestational age was 18.8 weeks (range 13-32). Karyotypes were available in 128 fetuses. The overall prevalence of trisomy 21 was 11 of 377 (2.9%). The mean (± standard deviation) iliac wing angle in the normal fetuses was 68.2 degrees (±15.4 degrees) and 98.5 degrees (±11.3 degrees) in fetuses with trisomy 21 (P < .001). Using an ROC-derived absolute cutoff of 90 degrees, an abnormal iliac wing angle had sensitivity of 90.9% (ten of 11), specificity of 94.50/. (346 of 366), NPV of 99.7% (346 of 347), and PPV of 33.3% (ten of 30) to detect trisomy 21. Conclusion: Fetuses with trisomy 21 have greater iliac wing angles than do normal fetuses. Using an ROC-derived absolute cutoff of 90 degrees, we could detect 90.9% of fetuses with trisomy 21 with a PPV of 33% in our high-risk population. These findings suggest that iliac wing angle is a useful marker in antenatal screening for trisomy 21. (Obstet GynecoI1997i89:734-7. © 1997 by The American College of Obstetricians and Gynecologists.)
Sonographic markers associated with fetal karyotype abnormalities were limited initially to the detection of major structural malformations. Over the last decade, From the Departments of Obstetrics and Gynecology, Divisions of Maternal-Fetal Medicine, Oakwood Hospital and Medical Center, Dear• born, Michigan, and the University of Connecticut Health Center, Farmington, Connecticut.
734 0029-7844/97/$17.00
PII 50029-7844(97)00089-6
however, more subtle ultrasound markers have been identified in the second trimester of pregnancy to identify a fetus at increased risk for a karyotype abnor• mality. These subtle ultrasound markers include con• genital heart defects, increased nuchal skin thickness, abnormally short or low-set fetal ears, shortened prox• imal long bones (ie, femur and humerus), increased brightness or echogenicity of the fetal bowel, a two• vessel umbilical cord, clubbed feet, pyelectasis, hypo• plasia of the mid-phalanx of the fifth finger, or an increased space between the first and second toes. 1- 3 Trisomy 21 is the most common karyotype abnormal• ity found in newborns, approximately 1:800 in the general population. 4 Up to 80% of newborns with trisomy 21 have abnormalities of the pelvic bones. These include small acetabular angles, elongated taper• ing ischia, and bilateral widening and flaring of the iliac wingS. 5,6 Although many of these pelvic bone abnor• malities have been documented in the neonatal period, their presence and diagnostic utility in the antenatal period has not yet been investigated. Our objective was to evaluate the efficacy of the ultrasonic measurement of fetal iliac wing angle as a marker for trisomy 21 in the second trimester of pregnancy.
Materials and Methods We identified women with singleton gestations who were referred for a targeted ultrasound or genetic amniocentesis from January 1, 1995, to June 30, 1995. Fetuses with previously determined karyotype abnor• malities were excluded. Collected data included mater• nal age, referral indications, gestational age at time of ultrasound, karyotype results, and iliac wing angle measurement. Using an axial (transverse) view of the fetal pelvis, including the sacral spine, the angle between the right and left iliac wings was measured on-screen (Figure 1), and designated the "iliac wing angle." To ensure a true
Obstetrics & Gynecology
Table 1. Indications for Ultrasound Referral Indication
n
(%)
Advanced maternal age Abnormal triple screen Trisomy 21 risk> 1:270 MSAFP > 2.0 MoM Uncertain dates Family history of congenital anomaly Medical complication of pregnancy Medication/teratogen exposure Other
159
(42.2)
44 27
01.7)
59 29 28 19 12
05.6)
m
Th~
(7.2) (7.7)
(7.4)
(5.0) (3.2) ~
MSAFP = maternal serum alpha-fetoprotein; MoM = multiples of the median.
Figure 1. Iliac wing angle determination from axial section through fetal pelvis.
axial section perpendicular to the sacral spine, the image was taken at a level where both iliac bones were at their longest dimension and of approximately equal length. This level generally includes the dome of the fetal bladder, and is always below the umbilical cord insertion. All angle measurements were performed by the same examiner (MOB) using a previously described on-screen method? All measurements were taken and calculated before knowledge of the outcome. The diagnosis of trisomy 21 was made by karyotype results from amniocentesis or by newborn examination for phenotypic evidence of Oown syndrome with karyotype confirmation. The sensitivity, specificity, positive (PPV), and negative predictive values (NPV) for the iliac wing angle measurement were calculated using multiple cutoff points, and a receiver operating characteristic (ROC) curve was generated to identify the optimum iliac wing angle measurement. Interobserver variation was evaluated in 20 consecutive normal fe• tuses by two observers who were blinded to each other's results. Each observer independently obtained an axial image of the fetal pelvis and measured the pelvic angle. The absolute values of pair differences [mean :!:: standard deviation (SO)} were calculated. Our statistical analysis included descriptive statistics, Stu• dent t test for continuous variables, and for categoric data, with P less than .05 deemed significant.
cases. A diagnosis of trisomy 21 was made in 11 (2.9%) of the 377 fetuses. The mean (±SO) iliac wing angle measurement in the normal fetuses was 68.2 degrees (±15.4 degrees) compared with 98.5 degrees (±11.3 degrees) in fetuses with trisomy 21 (P < .00l). The normal iliac wing angle measurement was independent of gestational age (r = 0.04). Figure 2 shows the angle measurement in a normal fetus. Interobserver variation was evaluated in 20 consecutive normal fetuses with a mean (±SD) variation of 4.4 degrees (±2.8 degrees) and absolute differences between matched pairs ranged between 0 degrees and 11 degrees. Using an ROC-derived optimum cutoff point of 90 degrees (Figure 3), ten of the 11 fetuses with trisomy 21 could be identified. The iliac wing angle measurement of 90 degrees or greater had a sensitivity of 90.9% (ten of 11) and a specificity of 94.5% (346 of 366). The PPV
x:
Results Our study population of 377 women had a mean (± SO; range) maternal age of 31.8 (±6.2; 15-44) years. The mean (±SD; range) gestational age of the fetuses was 18.8 (:!::2.4; 13-32) weeks. The referral indications are listed in Table 1. Amniocentesis was performed in 128
VOL. 89, NO.5, PART 1, MAY 1997
Figure 2. Axial section through fetal pelvis demonstrating measure• ment of iliac wing angle (IL = iliac wing; SP = spine).
Bork et al
Iliac Wing Angle
735
1.0 0.9 0.8 0.7
= = >-
~
til
c:
Ql
en
r::-
80'
73'
II'
116 ' 1111'
0.6 0.5 0.4 0.3 0.2 0.1
0.0 ......"""T'"-.-..-.."""T"".......,.......-.-..-,-"'T"""....,.---r-".-l 0.00.1 0.20.30.40.50.60.70.80.91.0
1-speclflclty Figure 3. Receiver operating characteristic curve for fetal iliac wing angle.
was 33.3% (ten of 30), and the NPV was 99.7% (346 of 347) in our population.
Discussion Caffey and Ross 5 were the first to diagnose changes in the pelvic bones of infants with Down syndrome. These findings have been confirmed by several other investi• gators. 8 - 10 The classic features in these reports are bilateral flattening of the lower edges of the ilia, elon• gated tapering ischia, small acetabular angles, and bilateral widening and flaring of the iliac wings. Our findings indicate that some of the characteristic skeletal abnormalities of the pelvis seen in newborns with trisomy 21 are present in the second trimester of pregnancy. We found that the iliac wing angle, which represents the "flaring of the iliac wings" described by Caffey and ROSS,6 is significantly increased in trisomy 21 fetuses when compared with normal ones. Little has been reported previously regarding prena• tal ultrasound evaluation of the fetal ilia. Two re• ports ll •12 of ultrasonographic measurement of fetal iliac bone length in normal fetuses have shown a linear relationship between fetal iliac bone length and gesta• tional age. Abuhamed et al13 have evaluated the utility of ultrasonographic iliac bone length in the identification of fetuses with Down syndrome in the second trimester of pregnancy. They have shown that iliac length in Down syndrome fetuses (longest bony dimension) is significantly longer than that of controls (P < .00l). In their study, 40% of trisomy 21 fetuses were identified by a measured-ta-expected iliac length ratio of at least 1.21. These results differ from our own study which mea-
736 Bark et
at Iliac Wing Angle
sures the angle described by the intersection of lines drawn through an axial cut of two iliac bones. The iliac bones are easily visualized, but technical difficulty in obtaining the proper image plane may occur when either the iliac wings are not of equal size (generally representing lateral obliquity with wings imaged at different heights) or if the intersection with the fetal anterior abdominal wall is below the dome of the bladder or at or above the umbilical cord insertion (suggesting anterior-posterior obliquity). To ensure uni• formity of measurements, we obtained axial sections through the fetal pelvis at the level of the upper sacral spine. This is approximately the same plane Abuhamed et al 13 used to measure the maximum length of the iliac wing. Although the angles in our study were electron• ically measured by calipers, the ROC-derived absolute cutoff of 90 degrees lends itself to visual discrimination of normal from abnormal. The prevalence of trisomy 21 in our study population of 2.9% (11 of 377) reflects our high-risk referral popu• lation, and is a limitation of our study. In a lower-risk population, given a lower prevalence, the false-positive rate would be higher. We extrapolated our data to a lower-risk population using Bayes theorem with the a priori risks derived from the patient's age or her triple screen. 3 For example, an abnormal iliac wing angle measurement increased the risk for trisomy 21 for a pregnant woman 35 years of age from 1 in 270 to 1 in 24. Conversely, until the a priori triple screen risk is less than 1 in 3500, an abnormal iliac wing angle at the time of a second trimester ultrasound will indicate a risk of trisomy 21 that is greater than 1 in 270. Further pra• spective studies in low-risk populations will be neces• sary to evaluate fully the utility of the iliac wing angle as a screening marker for trisomy 21. In our study, the iliac wing angle was 90.9% sensitive in detecting trisomy 21. Compared to the more com• monly recognized second-trimester ultrasonographic markers of aneuploidy, our findings suggest that iliac wing angle measurement will be among the most sensitive. Vintzileos and Egan3 reviewed the English language literature over a to-year period and reported the following mean (range) sensitivities: short femur 31% (13-70%), short humerus 33% (24-64%), and nu• chal fold thickening 34.1% (9-75%). In two of our 11 fetuses with trisomy 21, an increased iliac wing angle was the only ultrasound marker of aneuploidy. A third fetus had an isolated echogenic intracardiac focus of the left ventricle along with the increased iliac wing angle. The echogenic focus has also been described as a possible ultrasound marker for trisomy 21. 14 The re• maining eight fetuses had at least one of the more commonly recognized second-trimester ultrasono• graphic markers of aneuploidy. A priori knowledge by
Obstetrics & Gynecology
the ultrasonographer of these subtle anatomic changes suggestive of trisomy 21 could have introduced bias into the pelvic angle measurement; this may be a limitation of the study. Larger prospective studies in low-risk populations are required to assess accurately the utility of this marker.
References 1. Benacerraf BR, Neuberg 0, Bromley B, Frigoletto FD. Sonographic scoring index for prenatal detection of chromosomal abnormali• ties. J Ultrasound Med 1992;11:449-58. 2. Nyberg DA, Resta RG, Luthy DA, Hickok DE, Mahony B5, Hirsch JH. Prenatal sonographic findings of Down syndrome: Review of 94 cases. Obstet Gynecol 1990;76:370-7. 3. Vintzileos AM, Egan JFX. Adjusting the risk for trisomy 21 on the basis of second-trimester ultrasonography. Am J Obstet Gynecol 1995;172:837-44. 4. Thompson M, McInnes RR, Willard HF. Clinical cytogenetics: General principles and autosomal abnormalities. In: Thompson JS, Thompson MW, eds. Genetics in medicine. 5th ed. Philadelphia: WB Saunders, 1991:201-29. 5. Caffey J, Ross S. Mongolism (mongoloid deficiency) during early infancy: Some newly recognized diagnostic changes in the pelvic bones. Pediatrics 1956;17:642-51. 6. Caffey J, Ross S. Pelvic bones in infantile mongoloidism: Roent• genographic features. Am J RoentgenoI1958;80:458-67. 7. Bork MD, Egan JFX, Diana OJ, Scorza WE, Fabbri EL, Feeney LD, et al. A new method for on-screen ultrasound determination of fetal cardiac axis. Am J Obstet Gynecoll995;173:1192-5.
VOL. 89, NO.5, PART 1, MAY 1997
8. Kaufmann HJ, Taillard WF. Pelvic abnormalities in mongols. BMJ 1961;1:948-9. 9. Hall B. Mongolism in newborns: A clinical and cytogenic study. Acta Paediatr Scand Suppl 1964;154:41-8. 10. Mortensson W, Hall B. Abnormal pelvis in newborn infants with Down's syndrome. Acta RadioI1972;12:847-55. 11. Apuzzio n, Adhate A, Ganesh V, Leo MY, Holland BK. Prenatal ultrasonographic fetal iliac bone measurement: Correlation with gestational age. J Reprod Med 1992;37:348-50. 12. Jaffe R, Santolaya J, Warsof SL. Sonographic measurement of the fetal iliac bone: Growth and relation to femur length in assessment of gestational age. Am J Perinatol 1993;10:105-8. 13. Abuhamed AZ, Kolm P, Marl G, Slotnick RN, Evans AT. Ultra• sonographic fetal iliac length measurement in the screening for Down syndrome. Am J Obstet Gynecol 1994;171:1063-7. 14. Bromley B, Lieberman E, Laboda L, Benacerraf BR. Echogenic intracardiac focus: A sonographic sign for fetal Down syndrome. Obstet Gynecoll995;86:998-1001.
Reprints are not available.
Received November 7, 1996. Received in revised form January 10, 1997. Accepted January 28, 1997.
Copyright © 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Bork et al Iliac Wing Angle
737