ELSEVIER
European Journal of Obstetrics & Gynecology
and Reproductive Biology 72 (1997) 27-29
Interobserver variability of midtrimester fetal nuchal thickness
Antoni Borrell a'*, Dolors Costa b, Ruben D. Delgado a, Josep M. Martinez ", Carme Borrell c,
Albert Fortuny ~
~'Prenatal Diagnosis Unit, Department of Obstetrics and Gynecology, Hospital Clinic, University oJ Barcelona Medical School, Villarroel 170,
08036 Barcelona, Catalonia, Spain
bGenetics Service, Hospital Clinic, Barcelona, Catalonia, Spain
CMunicipal Health Institute, Barcelona, Catalonia, Spain
Received 11 July 1996; revised 18 September 1996; accepted 28 October 1996
Abstract
Objective: To assess the interobserver variability of ultrasound measured fetal nuchal thickness (NT). Study design: NT was
measured in 103 consecutive early midtrimester pregnancies (13-18 weeks) by three ultrasonographists with different levels of
expertise, in our Prenatal Diagnosis Unit. Measurements were obtained prior to amniocentesis, blindly in relation to other
examiners' values. The Lin's concordance correlation coefficient and the Kappa statistics were used to assess concordance between
observed values. Results: Mean differences between pairs of observers ranged from 0.44 to 0.57 mm. A higher concordance
correlation between measurements was obtained between the more experienced observers (r=0.77 versus r= 0.51). When
assigning cases with a NT .>_5 mm to a group at risk for Down's syndrome, a high concordance between experienced
ultrasonographists was observed (98% agreement; K = 0.90). Conclusion: Interobserver variability in NT was acceptable between
experienced ultrasonographists. © 1997 Elsevier Science Ireland Ltd.
Keywords: Down's syndrome screening; Interobserver variability; Nuchal thickness; Ultrasonography
1. Introduction
Fetal nuchal skinfold thickness (NT) measurement has
been suggested as a useful ultrasound marker to screen
for Down's syndrome in the second trimester of pregnancy
[1]. Initial findings pointed out that 43% of the fetuses
affected by Down's syndrome were detected, being only
0.1% of the euploid fetuses selected by the same criteria.
However, subsequent studies showed a wide range of
variability in the results. In prospective studies sensitivity
for Down's syndrome ranges from 16 to 75%, for false
positive rates from 0.3 to 8.5% [2-7]. In our center, a 78%
sensitivity was achieved in a series of 1543 consecutive
cases undergoing midtrimester amniocentesis, for a 2.1%
false positive rate [8].
* Corresponding author: Tel: +34 3 2275400, fax: + 34 3
2275454.
One of the main reasons suggested to explain such
differences in the results, was interobserver variability.
To our knowledge, few studies have attempted to clar-
ify this issue [3,9]. In our Prenatal Diagnosis Unit, 103
consecutive cases were assessed by ultrasonographists
with three different levels of expertise in order to estab-
lish the interobserver variability of NT.
2. Material and methods
Fetuses (n = 103) corresponding to consecutive cases
of amniocentesis performed in our Unit during April
and May 1994 were studied. Previous to the procedure,
NT was measured in each fetus by three observers with
different levels of expertise: high (Observer A), interme-
diate (either Observer BI or B2), and low (C). Levels of
0301-2115/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved
PII S0301-21 15(96)02659-0