Reliability of estimating ductions in thyroid eye disease: an International Thyroid Eye Disease Society multicenter study.

PURPOSE: To determine intraobserver and interobserver reliability of 3 clinical techniques for measuring ocular ductions in patients with thyroid eye disease and to compare these with an established method using a Goldmann perimeter. Our secondary goals were to compare the clinical methods for test duration, ease of learning and performance, and whether these differed between novices and experts.

DESIGN

Multicenter, prospective, comparative trial.

PARTICIPANTS

We recruited 38 patients with thyroid orbitopathy and reduced ocular motility from 7 academic centers.

METHODS

At each center, 2 novices and 2 experts measured the ocular ductions of each eye of each patient in 4 directions (0°, 90°, 180°, and 270°) using 3 techniques (best guess [BG], light reflex [LR], and a ruler measuring [RU] technique). Test duration and clinicians' feedback were recorded. A subjective and objective measurement for ocular ductions using a Goldmann perimeter was performed for each subject by a trained technician. The 3 clinical measurements and the perimetry measurements were performed twice, separated by ≥1 hour.

MAIN OUTCOME MEASURES: We measured the intraobserver and interobserver reliability of the 3 clinical techniques and intraobserver reliability of Goldmann perimeter. Clinical testing reliability was compared between experts and novices. We also examined test duration and clinician feedback.

RESULTS

The LR technique had significantly better intraobserver and interobserver repeatability compared with the BG or RU clinical measurements and statistically was equivalent to the gold-standard perimetric technique. Reliability was constant regardless of the amount of restriction in ocular movement. There was no difference between reliability values for experts and novices. The LR and BG techniques were significantly faster than the RU or perimetry techniques and were considered easiest to learn and perform, but clinicians had most confidence in the LR technique results.

CONCLUSIONS

The LR technique for measuring ocular ductions in thyroid orbitopathy is more reliable than other clinical techniques and as reliable as the established technique using the perimeter. However, unlike the latter method, it is easier to learn and perform by both novices and experts, is significantly faster, and can be performed by the clinician without machinery or a trained technician.