Dear Editor

We read with interest the article by Ebner et al [1] investigating the efficacy of periocular triamcinolone for the treatment of Thyroid Associated Ophthalmopathy (TAO) and the presence of ocular or systemic adverse effects also previously published in 2001 [2]. The study used patients with TAO of less than 6 months duration previously untreated, and assumed that the activity of the TAO would be equal.

However large differences are apparent in both the demographics (a 78 year old presenting with new onset TAO, and the use of an 11 year old child both raised interesting scientific and ethical questions.) and the area of diplopia at baseline suggesting baseline activity was not equal between groups. No mention was made of either smoking or systemic medications used, both of which are factors affecting the activity of TAO, and the subsequent measurement of the potential systemic adverse effects of intraorbital triamcinolone. The main measure of the local effect of triamcinolone was the area of single vision obtained on a Goldmann perimeter, but the use of a 2-IV size light and the measurement of “summation of angular points” is unclear.

The standard Goldmann nomenclature generally used designates a Roman numeral for target size, and a combination of a number and letter to signify target brightness and intensity. A 2-IV suggests a 4.51mm diameter light was used but the brightness and intensity measure are not evident. The optimal size is a spot subtending an angle of 2° to help discern the point of diplopia; a size IV only subtends an angle of 0.86°and maybe to small to discriminate accurately. The summation of angular points method to assess the area of single vision on the graph obtained does not seem logical. Why not calculate the area within the graph? No history of either prior strabismus or suppression seems to have been elicited from the patients that would have a large effect on the results.

Both graphs shown demonstrate a large number of point readings taken superiorly with only relatively few inferiorly, making the graph area a less reliable measure. A more suitable method may have been a baseline independent orthoptic assessment with serial Lees screen or Lister perimeter measurements, and the adoption of Bagolini’s glasses to minimise suppression. The effects of triamcinolone locally were measured with exophthalmometry, optic disc appearance (normal, papilloedema, atrophy), ocular motility assessment and intraocular pressure (IOP). The authors demonstrated in fig.3 an example of improved cosmesis obtained but failed to measure this during the study, with perhaps serial photographs [4]. Exophthalmometry values were not published, but any change was stated to be not statistically significant.

The authors failed to explain why they thought the significant difference between the areas of no diplopia demonstrated between the treatment and control group were not supported by either a significant change in either proptosis or extra ocular muscle diameter (except superior rectus) both of which are good clinical indicators of TAO activity. Systemically the effect of triamcinolone was measured by recording the body weight, blood pressure and various blood tests at baseline, week 10 and week 24. Single point measurement of these variables is unlikely to provide statistically significant information especially when confounding factors such as systemic medications taken are unknown.

References

(1) Ebner R, Devoto M,Weil D et al. Treatment of thyroid associated ophthalmopathy with periocular injections of triamcinolone. Br J Ophthalmol 2004;88:1380-1386.

(2) Ebner R, Devoto M,Weil D et al. Tratamiento de la oftalmopatia asociada a distiroidismo con triamcinolona periocular . Arch Oftalmol Bs As 2001;76:55-66.

(3) Dickinson AJ, Perros P, :Controversies in the clinical evaluation of active thyroid associated thyroid orbitopathy. Clin Endocrinol 2001;55:283 -303.