Concordance of three methods for palpebral fissure length measurement in the assessment of fetal alcohol spectrum disorder

Abstract

Background: The assessment of individuals at risk of fetal alcohol spectrum disorders (FASD) includes the assessment of the craniofacial features that can result from prenatal alcohol exposure. The characteristic facial features of fetal alcohol syndrome (FAS) consist of short palpebral fissures, smooth or flattened philtrum, and thin vermilion border of the upper lip. There are various methods for measuring palpebral fissure lengths (PFLs) and it can be challenging for clinicians to obtain reproducibly accurate measurements. The development of the FAS Facial Photographic Analysis Software by the University of Washington FAS Diagnostic and Prevention Network (DPN) is one such means of improving the accuracy and reproducibility in these measurements.

Objectives: To assess concordance across three methods of PFL measurement: 1) a clear plastic handheld ruler, 2) blunt precision slide calipers, and 3) digital photometric photography (FAS Facial Photographic Analysis Software).

Methods: The PFLs of 50 children (referred to the Clinic for Alcohol and Drug Exposed Children, CADEC) at Children's Hospital in Winnipeg and 50 adults from the University of Manitoba Medical Class of 2008 were measured once by a single clinician, using each of the three methods. The frequency and magnitude of discordance was tabulated. No method served as a gold-standard.

Results: The PFLs ranged from 20 to 32 mm. The ruler and photometric measures were concordant in 42% of the subjects. When measures were discordant, half the ruler measures were larger and half were smaller. The caliper measure was concordant with the photometric and ruler measures on 18% and 24% of the subjects, respectively. When measures were discordant, the caliper measures were almost always larger than the photometric and ruler method (0.5 to 2.5 mm larger, 83% and 95% of the time, respectively). The presence of epicanthal folds did not appear to be a factor that contributed to discordance.

Conclusion: This study demonstrates the challenge in measuring the PFL, even when a single trained clinician is involved. Factors that can contribute to error include the subject's willingness to cooperate, ability to tolerate placement of the tool close enough to the eye to obtain an accurate measure, and precision of the tool. When controlling for the clinician performing the measurements and the quality of the photographs, the ruler and photometric measures were most concordant. The caliper measures tended to measure larger than the ruler and photometric measures.