Skip Navigation LinksHome > February 2007 - Volume 109 - Issue 2, Part 1 > Absent or Shortened Nasal Bone Length and the Detection of D...
Obstetrics & Gynecology:
doi: 10.1097/01.AOG.0000250903.17964.87
Original Research

Absent or Shortened Nasal Bone Length and the Detection of Down Syndrome in Second-Trimester Fetuses

Gianferrari, Elisa A. MD1; Benn, Peter A. DSc2; Dries, Lisa DO3; Brault, Kim RD MS3; Egan, James F. X. MD1; Zelop, Carolyn M. MD3

Free Access
Article Outline
Collapse Box

Author Information

From the 1Division of Maternal–Fetal Medicine, Department of Obstetrics and Gynecology, University of Connecticut Health Center, Farmington, Connecticut; 2Division of Human Genetics, Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut; and 3Department of Obstetrics and Gynecology, St. Francis Hospital and Medical Center, Hartford, Connecticut.

Corresponding author: Carolyn M. Zelop, MD, Department of Obstetrics and Gynecology, St. Francis Hospital and Medical Center, 114 Woodland Street, Hartford, CT 06105; e-mail:

Collapse Box


OBJECTIVE: To estimate the accuracy of evaluating nasal bone length, expressed as multiples of the median (MoM), for the detection of Down syndrome in second- trimester fetuses.

METHODS: Expected normal median nasal bone measurements were established for an initial cohort of women receiving fetal ultrasound examinations at 15–24 weeks of gestation. Nasal bone lengths were converted to MoM with adjustment for maternal race and ethnicity using whites as the referent group. Nasal bone MoM were compared in euploid and Down syndrome fetuses. The sensitivity and specificity were evaluated in this initial cohort and in a second cohort in which all ultrasound measurements were carried out prospectively.

RESULTS: For the combined data set, 10 of 21 affected pregnancies had an absence of the nasal bone (sensitivity 47.6%), but absence was noted in only 1 of 2,515 unaffected pregnancies (false-positive rate 0.04%). Using less than 0.80 MoM as a cutoff, the sensitivity was 20 of 21 (95.2%), and the false-positive rate was 185 of 2,515 (7.4%). Changing the cutoff to 0.75 MoM resulted in 18 of 21 (85.7%) sensitivity and 74 of 2,515 (2.9%) false-positive rate. Using medians derived from whites to calculate MoM for the entire population resulted in higher false-positive rates.

CONCLUSION: Nasal bone length expressed as MoM seems to be an useful ultrasound marker for Down syndrome in second-trimester fetuses with a high sensitivity and a low false-positive rate.


Langdon Down, in his report from the London Hospital in 1866, described the Down syndrome phenotype and characterized the “face as flat and the nose as small.”1 A small nose and the presence of epicanthal folds were subsequently shown to be consistent features of Down syndrome and are thought to be due to delayed development of the nasal bone.2 It was surmised that this finding might translate into the prenatal ultrasound finding of nasal bone hypoplasia or absence of the fetal nasal bone.

The first description of nasal bone hypoplasia or “absence of the nasal bone” was published by Sonek and Nicolaides,3 where three cases of trisomy 21 had either an absent nasal bone (two cases) or a nasal bone measurement at less than the 2.5 percentile (one case). This was followed by a large observational study where these investigators established normal ranges for the nasal bone length throughout gestation. Approximately 3,500 nasal bone measurements were made on unselected patients between 15 and 22 weeks of gestation to establish normal ranges.4

Because the size of the nasal bone in unaffected pregnancies is gestational age dependent, investigators have proposed different methods to define nasal bone hypoplasia with variable efficacy. Bromley et al5 proposed that absent fetal nasal bone or increased biparietal diameter/nasal bone ratio would identify Down syndrome in second-trimester fetuses in a high-risk population. In their population, biparietal diameter/nasal bone of 10 or more identified 81% of the Down syndrome fetuses with an 11% false-positive rate. Using the same criterion, Tran et al6 also obtained an 81% detection rate but the false-positive rate was 44%. Odibo et al7 reported that use of biparietal diameter/nasal bone more than 11 provided a sensitivity of 59% and a false-positive rate of 15% for the detection of Down syndrome. In their study, absence of the nasal bone identified 23% of Down syndrome cases with a 1% false-positive rate in second-trimester fetuses. Defining nasal bone hypoplasia as less than 2.5mm in second-trimester fetuses, Cicero et al2 detected shortened nasal bone length in 61.8% of Down syndrome fetuses, producing a likelihood ratio of 50.5. Using a fetal nasal bone length less than the fifth percentile as a screening cutoff, Bunduki et al8 reported a sensitivity of 59.1% and a 5.1% false-positive rate when screening for Down syndrome. Alternatively, nasal bone length can be converted to multiples of the median (MoM) allowing for differences in gestational age at the time of measurement. Recently, Maymon et al9 combined prenasal thickness with nasal bone length, both expressed as MoM, reporting a 70% detection rate (compared with a 43% rate using nasal bone length alone) for a 5% false-positive rate.

We sought to further estimate the utility of absent and small nasal bone length expressed as MoM for Down syndrome screening in our population.

Back to Top | Article Outline


The fetal nasal bone was identified as a discrete echogenic line in a midsagittal plane of the fetal profile and was measured from the base of the nose closest to the frontal bone to the most distal aspect of ossification (Fig. 1). The nasal bone measurement was carried out as described by Sonek et al3 and required that the ideal angle of insonation approximate to either 45 degrees or 135 degrees. Tabulated demographic characteristics for each case included maternal age at delivery, gestational age as determined by ultrasound, and maternal race and ethnicity according to the following categories as identified by the patient: white, Hispanic, Asian, African-American, and “Other.”

Fig. 1
Fig. 1
Image Tools

To establish the expected normal distribution of nasal bone lengths for the second trimester, we reviewed the fetal nasal bone measurements in a series of 732 women of known race and ethnicity referred for second-trimester ultrasound evaluation between May 2003 and July 2004. The size of this cohort was based on the approximate number of cases generally used to establish population median values in maternal serum screening. Details of this population have previously been reported.10 Cases where the pregnancy was affected with Down syndrome or other significant chromosome abnormality were excluded. To express each nasal bone measurement as an MoM, 498 measurements from white women were sorted by gestational week and the median log10 transformed nasal bone values were calculated for each week. A regression curve was then established that expressed the median nasal bone length as a function of ultrasound gestational age. Each patient's fetal nasal bone measurement was then expressed as a multiple of the normal median using the expected gestational day-specific median determined from the regression curve. For race or ethnicity other than white, normal median values were derived from the white data and applying a multiplicative correction factor (1.0412 for African-Americans, 1.0522 for Hispanics, 0.9480 for Asians and 1.0376 for “Others”). These correction factors were derived by dividing the median MoM for whites by the median MoM for each race.10

In the first cohort, the Down syndrome cases were measured retrospectively from archived images in our ultrasound database after the cytogenetic diagnoses were made. In those cases with a measurable nasal bone, the nasal bone length was converted into race or ethnicity–adjusted MoM using the same methods.

To further assess the usefulness of nasal bone length as a marker for fetal Down syndrome, we carried out additional analyses on a second cohort of 1,794 women referred for second-trimester ultrasound between July 2004 and May 2006. In this second series, all ultrasound evaluations were prospective. Subsequent pregnancy outcome information indicated that 1,783 pregnancies were unaffected and 11 had fetal Down syndrome. For this second cohort, the conversion of nasal bone measurements to MoM was based on the gestational age–specific regression formula and race correction factors established in for the first cohort.

We determined the proportion of affected and unaffected cases with absence of a nasal bone, length less than 0.80 MoM, or length less than 0.75 MoM for the two cohorts considered separately and also for a combined data set. Sensitivity, false-positive rates (1-specificity), positive predictive values, and likelihood ratios with 95% confidence intervals (CIs) were generated. We also evaluated the efficiency of nasal bone evaluation when medians derived for whites were applied to the entire population.

To assess the results using biparietal diameter/nasal bone ratios, we calculated this ratio for each patient with a correction for race or ethnicity established using the same approach as described above for MoM. The correction factors were 0.9733 for African-Americans, 0.9728 for Hispanics, 1.0500 for Asians, and 0.9394 for “Others.” Detection rates and false-positive rates were calculated for biparietal diameter/nasal bone ratio cutoffs of more than 8, 9, and 10.

This study was approved by the institutional review board of the St. Francis hospital.

Back to Top | Article Outline


Table 1 summarizes race or ethnicity data for cohort 1, cohort 2, and the combined data set. Mean maternal age at delivery was 30.9 years in unaffected pregnancies 33.8 years in affected pregnancies.

Table 1
Table 1
Image Tools

Figure 2 shows the relationship between median nasal bone length and gestational age in unaffected pregnancies derived from cohort 1 women. The regression curve was described by: log10 {median [nasal bone MoM}]=0.03925 × {gestational age [weeks}] + 0.21012; (r 2=0.97). Also shown in Figure 2 are the values of nasal bone length corresponding to the 0.80 and 0.75 MoM cutoffs used in this study. After converting all cohort 1 unaffected pregnancy data into race or ethnicity–adjusted MoM, the median MoM was 0.989, which was consistent with the expected value of 1.000.

Fig. 2
Fig. 2
Image Tools

For cohort 2 unaffected pregnancy data, the median MoM was 0.988. This median value provided a check on the conversion into MoM for cohort 2 and helped establish the validity of using a regression formula derived from cohort 1 data to calculate MoM in the subsequent cohort 2 patients. For unaffected pregnancies, there was no significant difference between the MoM values of retrospective data and prospective data (unpaired t test, P=.12).

Table 2 summarizes the sensitivities and false-positive rates for absence of a nasal bone, length less than 0.80 MoM and length less than 0.75 MoM. Results for the latter two cutoffs include cases where there was no measurable nasal bone. For both the detection rates and false-positive rates, there was no significant difference between the results for retrospective and prospective data (Fisher exact test, P>.05), and results of the two cohorts were therefore combined to provide the most robust estimates. For the combined data, absence of a nasal bone was noted in only 1 of 2,515 unaffected pregnancies. Absence of a nasal bone had a sensitivity of 47.6% and specificity 99.96%. Absence of a nasal bone was therefore associated with a very high probability of Down syndrome with a positive predictive value of 90.9%. Using the 0.80 MoM cutoff, the sensitivity was increased to 95.2% with a specificity of 92.72% and a positive predictive value of 9.85%. The performance based on the 0.75 MoM cutoff was: sensitivity 85.7%, specificity 97.13%, and positive predictive value 17.6%. The false-positive rate (1-specificity) was therefore strongly dependent on the choice of cutoff increasing from 74 of 2,515 (2.9%) to 185 of 2,515 (7.4%) when the cutoff was changed from 0.75 MoM to 0.80 MoM. Using the set of medians derived from white women to evaluate the entire group did not affect the detection rates. However, the false-positive rates were increased to 3.3% and 8.8% for the 0.75 MoM and 0.80 MoM cutoffs, respectively. Table 2 also summarizes the positive likelihood ratios (multiplicative factor associated with a nasal bone hypoplasia) and negative likelihood ratios (multiplicative factor associated with nasal bone length greater than the cutoff) for the combined data with race or ethnicity adjustments.

Table 2
Table 2
Image Tools

Maternal age at delivery seemed to be independent of nasal bone MoM in both unaffected (Pearson correlation coefficient, r=–0.03, P=.25) and affected (r=0.40 P=.22) fetuses. There was no difference in maternal age at delivery in Down syndrome fetuses with and without measurable nasal bone lengths (unpaired t test, P=.33).

Using race or ethnicity–adjusted biparietal diameter/nasal bone ratios also preferentially identified Down syndrome–affected pregnancies. For the combined data set (and including those cases where there was an absence of nasal bone), a biparietal diameter/nasal bone cutoff of more than 8 had a sensitivity of 100% (CI 84.5–100.0%) and a specificity of 85.81% (CI 84.4–87.1%). Increasing the cutoff to more than 9 reduced the sensitivity to 90.5% (CI 71.1–97.4%) and increased the specificity to 96.8% (CI 96.1–97.4%). A more than 10 cutoff gave a sensitivity of 81% (CI 60.0–92.3%) and specificity of 99.4% (CI 99.0–99.6%).

For a fixed 5% false-positive rate, use of nasal bone MoM identified 19 of the 21 affected pregnancies (sensitivity 90.5%, CI 71.1–97.5%), whereas use of biparietal diameter/nasal bone ratio identified 20 affected pregnancies (sensitivity 95.2%, CI 77.3–99.2%). The difference in cases identified was not significantly different (McNemar test, P=1.0).

Back to Top | Article Outline


In this study we have confirmed that second-trimester ultrasound evidence for absence of a nasal bone is a useful marker for Down syndrome, allowing for the identification of nearly half of all affected pregnancies with a very low false-positive rate. Furthermore, in those affected pregnancies with a measurable nasal bone, there is hypoplasia relative to gestational age-matched normal pregnancies, which provides potential identification of additional cases.

Although our data offers potential clinical usefulness, our study has several limitations that must be recognized. We acquired one half of our Down syndrome cases through retrospective review of previously archived images that may introduce bias into our data. This bias includes influence in the placement of calipers when the diagnosis is already known, the preferential inclusion of affected pregnancies ascertained through maternal serum screening or other ultrasonographic criteria, and possible viability bias of affected pregnancies. In both cohorts, the prevalence of Down syndrome–affected cases was very high and the derived likelihood ratios are not necessarily applicable to low-risk patients due to these biases. We also do not know in what percentage of cases or controls we were unable to obtain a reliable image in which the fetal nasal bone could be assessed. Interobserver and intraobserver variation was also not assessed.

Our results are broadly consistent with previous studies that evaluated nasal bone length in the second trimester. These studies have shown that some second trimester Down syndrome fetuses do have an apparent absence of the nasal bone whereas others show hypoplasia relative to unaffected fetuses.2–9

Both the use of MoM and biparietal diameter/nasal bone ratios seem to provide practical approaches to screening for fetal Down syndrome, because we were unable to demonstrate any significant difference between the two methods. The various studies that have used biparietal diameter/nasal bone ratios have reported widely different sensitivities and specificities for same cutoff (5–7, this study). We were also able to show that the false-positive rate differs markedly when the cutoff was changed by a relatively small amount, from 0.75 MoM to 0.80 MoM. It is therefore likely that center-specific normal population measurements and a high degree of standardization in measurement technique may be required whichever method is used. In the absence of any difference in performance, the use of MoM may be preferred over biparietal diameter/nasal bone ratios because departures from normal (1 MoM) are more readily recognized and the format is consistent with many other screening markers.

One important difference in our approach compared with previous studies involving nasal bone measurement was the use of an adjustment for race or ethnicity. These minor adjustments for race or ethnicity did alter the performance of nasal bone measurement as a screening test. Similar to the procedure used in maternal serum screening, the adjustments for race or ethnicity were based on women's self-assignment of their group and further refinement might be possible for mixed parentage or for specific ethnic subgroups that were included in the broad classifications used in this study.

We found that nasal bone length MoM appeared to be independent of maternal age. This result is consistent with the report of Tran et al.6 This suggests that likelihood ratios could be used to modify maternal age–specific risks for Down syndrome. It remains important to establish whether there is correlation between nasal bone length and other ultrasonographic or serum markers before routinely using likelihood ratios to modify risks based on these additional criteria.

In the absence of a large, fully prospective study that establishes standardized nasal bone measurements as a screening test for Down syndrome, we are hesitant to advocate the modification of individual women's risks on the basis of nasal bone measurements that might suggest mild or even moderate hypoplasia. However, we do suggest that careful second-trimester scanning of the midsagittal plane of the fetal facial profile specifically looking for presence or absence of a nasal bone be carried out. Previous studies together with our results indicate that the unambiguous absence of nasal bone is an indication that Down syndrome may be present.

Back to Top | Article Outline


1. Down JL. Observations on an ethnic classification of idiots. London Hospital. Clinical Lectures and Reports 1866;3:259–62.

2. Cicero S, Sonek JD, McKenna DS, Croom CS, Johnson L, Nicolaides KH. Nasal bone hypoplasia in Trisomy 21 at 15-22 weeks' gestation. Ultrasound Obstet Gynecol 2003;21:15–8.

3. Sonek JD, Nicolaides KH. Prenatal ultrasonographic diagnosis of nasal bone abnormalities in three fetuses with Down syndrome. Am J Obstet Gynecol 2002;186:139–41.

4. Sonek JD, McKenna D, Webb D, Croom CS, Nicolaides KH. Nasal bone length throughout gestation: normal ranges based on 3537 fetal ultrasound measurements. Ultrasound Obstet Gynecol 2003;21:152–5.

5. Bromley B, Lieberman E, Shipp TD, Benacerraf B. Fetal nose bone length: a marker for Down syndrome in the second trimester. J Ultrasound Med 2002;21:1387–94.

6. Tran LT, Carr DB, Mitsumori LM, Uhrich SB, Shields LE. Second-trimester biparental diameter/nasal bone length ratio is an independent predictor of trisomy 21. J Ultrasound Med 2005;24:805–10.

7. Odibo AO, Sehdev HM, Sproat L, Parra C, Odibo L, Dunn L, et al. Evaluating the efficiency of using second-trimester nasal bone hypoplasia as a single or a combined marker for fetal aneuploidy. J Ultrasound Med 2006;25:437–41.

8. Bunduki V, Ruano R, Miguelez J, Yoshizaki CT, Kahhale S, Zugaib M. Fetal nasal bone length: reference range and clinical application in ultrasound screening for trisomy 21. Ultrasound Obstet Gynecol 2003;21:156–60.

9. Maymon R, Levinsohn-Tavor O, Cuckle H, Tovbin Y, Dreazen E, Wiener Y, et al. Second trimester ultrasound prenasal thickness combined with nasal bone length: a new method of Down syndrome screening. Prenat Diagn 2005;25:906–11.

10. Zelop CM, Milewski E, Brault K, Benn P, Borgida AF, Egan JF. Variation of fetal nasal bone length in second trimester fetuses according to race and ethnicity. J Ultrasound Med 2005;24:1487–9.

© 2007 The American College of Obstetricians and Gynecologists


Article Tools