Volume 197, Issue 4 , Pages 335-336, October 2007
Nasal bone in screening for trisomy 21: defining hypoplasia
Article Outline
For well over a century, it has been recognized that 1 of the common dysmorphic features of individuals with Down syndrome is a comparatively small nose.1 Anthropometric studies have shown that 50% of these individuals have an abnormally short nasal root, which provides objective support for this observation.2 Radiologic studies in terminated second-trimester fetuses with trisomy 21 have demonstrated a lack of ossification of the nasal bone (NB) in 33% of the cases and an abnormally short NB in 23% of the cases.2
See related article, page 361
Over the past 7 years, accumulated literature has shown examination of the NB with prenatal ultrasound scanning to be useful in screening for trisomy 21 in both the first and second trimesters. The difference in the examinations is that only the presence or absence of the NB is noted in the first trimester; in the second trimester, NB length (NBL) measurement has also been shown to be informative.2
Combined data from studies that were performed between 11-13+6 weeks of gestation and where a standardized approach to NB evaluation was used to yield a total of 35,213 fetuses that were examined. The NB was absent in 66.8% of fetuses with trisomy 21 and in 0.9% of euploid fetuses.2, 3
A review of second-trimester studies (3014 fetuses at 15-26 weeks of gestation) showed a prevalence of NB absence in 37% of fetuses with trisomy 21 and 0.9% of chromosomally normal fetuses.2 Data from studies that looked at the prevalence of NB hypoplasia found it to be present in 48.2% of trisomy 21 fetuses and in 2.4% of chromosomally normal fetuses (n = 5726).2 The prevalence of either NB absence or hypoplasia was found to be 60% in trisomy 21 fetuses and 1.4% in chromosomally normal fetuses (n = 2240).2 Evaluation of the NB improves the performance of the second-trimester genetic ultrasound scans in a statistically significant way.4
The definition of NB absence is clear, and the diagnosis is relatively easy to make in the second trimester. However, a number of different definitions have been used in the literature to describe NB hypoplasia. They fall into 3 general categories: NBL adjusted for gestational age by the use of biparietal diameter (BPD) to NBL ratio (BPD/NBL), a single measurement cut-off, and defining hypoplasia on the basis of gestational age–adjusted NBL distributions in normal fetuses.
The use of BPD/NBL ratios was introduced by Bromley et al,5 who studied 239 euploid fetuses and 16 fetuses with Down syndrome between 16 and 23 weeks of gestation. They reported that 37% of the Down syndrome fetuses had an absent NB. For a BPD/NBL of 10, the detection rate was 81%, for a screen positive rate of 11%. The NBL in the euploid fetuses increased linearly, whereas NBL in the fetuses with Down syndrome remained relatively constant across the gestational ages that were studied.
Cicero et al6 defined NB hypoplasia as an NBL of ≤2.5 mm. The prevalence of absent NB was 32.4%. The prevalence of either NB hypoplasia or absence in the 34 fetuses with Down syndrome was 61.8%, with a screen positive rate of 1.2%. The prevalence of NB hypoplasia was highest in persons of Afro-Caribbean descent (8.8%).
Finally, a number of studies have defined NB hypoplasia as an NBL below a certain percentile (5th or 2.5th), which results in detection rates that vary from 48.4%-100%.2
In the current issue of the Journal, Odibo et al7 provide additional evidence that prenatal ultrasound evaluation of the NB is useful to assess the risk of trisomy 21. The authors successfully examined the NB in 3197 of 3634 fetuses (88% of the cases) between 15 and 23 weeks of gestation. These data were used to generate regressed means for each gestational week. For fetuses of ethnicities other than white, correction factors were used. Of the 23 cases trisomy 21, 6 cases (37%) had an absent NB; in the remaining 17 cases, NBL was measured.
The authors first compared the sensitivities and specificities at the point at which the screening performance of the multiples of the median (MoM) and BPD/NBL approaches was deemed optimal (MoM, <0.75; or BPD/NBL, >11). The sensitivities were 49% and 61%, respectively. The specificities were 92% and 84%, respectively. The sensitivity of the MoM approach was lower than the BPD/NBL approach, but the difference was not statistically significant. However, the MoM approach was found to have a statistically higher specificity.
The authors also compared the 2 approaches by fixing either the sensitivity or the screen positive rate. For a fixed sensitivity of 40%, the screen positive rates were 1% and 5.5% for the MoM and the BPD/NBL approaches, respectively. For a fixed 5% screen positive rate, the detection rates between the 2 approaches were similar at 39% and 43%, respectively.
Certain data that would be helpful to evaluate these results are not available in the article. Comparison of the distributions of values that were obtained in fetuses of various ethnicities to determine the need for the use of correction factors, the gestational age at which trisomy 21 fetuses were examined, the actual NBL measurements in the trisomy 21 fetuses, and a discussion regarding the fact that the NBL normal ranges in this study are generally lower than those that have been published previously are not presented.
Nonetheless, the data presented by Odibo et al indicate that the MoM approach is superior to the BPD/NBL approach. There are theoretic reasons to suggest that BPD/NBL ratios may not provide optimal results. First, the ratio method uses 2 fetal measurements and therefore 2 potential sources of error. Second, BPD/NBL ratios may not be constant as gestational ages change. Generally, when a ratio is created from a marker and a component of biometry, the resultant values tend to change with gestation because the relationship between biometry and gestation is often curvilinear rather than linear. Third, the use of the same ratio threshold across gestational age does not take into consideration that the discriminatory power of the marker may be gestational age–related.
An additional benefit to the MoM approach is that it not only serves as a way to define NB hypoplasia but also allows generation of likelihood ratios (LR) for various NBL measurements. Using LR’s to modify an a priori risk is the most efficient way to use continuous variable markers, especially if they are gestational age-specific, in screening for trisomy 21.
An important unresolved issue is whether efficiency of screening by NB evaluation changes during the first half of the second trimester. There is evidence that NBL in fetuses with trisomy 21 does not increase in the same fashion as in euploid fetuses. Therefore, it is possible that efficiency of screening with NBL during the first half of the second trimester improves with advancing gestational age. At present, we favor the MoM approach in our center.
References
- . Observations on an ethnic classification of idiots: clinical lectures and reports. London Hospital. 1866;3:259–262
- . Nasal bone assessment in prenatal screening for trisomy 21. Am J Obstet Gynecol. 2006;195:1219–1230
- . Nasal bone in first-trimester screening for trisomy 21. Am J Obstet Gynecol. 2006;195:109–114
- . Second-trimester biparietal diameter/nasal bone length ratio as an independent predictor of trisomy 21. J Ultrasound Med. 2005;24:805–810
- . Fetal nasal bone length: a marker for Down syndrome in the second trimester. J Ultrasound Med. 2002;21:1387–1394
- . Nasal bone hypoplasia in fetuses with Trisomy 21 at 15-22 weeks’ gestation. Ultrasound Obstet Gynecol. 2003;21:15–18
- . Defining nasal bone hypoplasia in second-trimester Down syndrome screening: does the use of multiples of the median (MoM) improve screening efficacy?. Am J Obstet Gynecol. 2007;(in press).
PII: S0002-9378(07)00998-2
doi:10.1016/j.ajog.2007.08.032
© 2007 Mosby, Inc. All rights reserved.
Refers to article:
- Defining nasal bone hypoplasia in second-trimester Down syndrome screening: does the use of multiples of the median improve screening efficacy?
Volume 197, Issue 4 , Pages 335-336, October 2007
