To develop and evaluate a novel bleb grading scheme for clinical and photographic evaluation.
A system for grading bleb photographs using widely applicable parameters was designed, and reference color photographs printed. A prospective masked agreement study was undertaken comparing slit lamp examination with mono and stereo photographs; 36 eyes of 28 patients with previous glaucoma surgery were graded according to defined parameters on a 1 to 10 scale clinically at the slit lamp by four ophthalmologists and two optometrists. Standardized stereo and mono photographs of the blebs were taken on the same day. The photographs were graded at least one week later in a masked fashion by the same observers, with grading of mono and stereo photographs also separated by one week. Analysis was performed to determine the variability and agreement between slit lamp results and photographic results, and to identify the presence of systematic bias.
High levels of agreement were found between slit lamp and both stereo and mono photographs for vascularity indices, bleb wall thickness, and bleb elevation. Lower levels of agreement were found for the relative components of demarcated versus diffuse areas of the bleb, and for the total width of the bleb. The interquartile range for the median difference between slit lamp and photograph grading was −1.0 to 1.0 for all criteria except diffuse component (−2.0 to 2.0), and the median difference for all scores was 0.0. The median interobserver difference for all criteria was 0.0; the quartile range for all scores was between −0.5 and 1.0 except for diffuse component and width assessments whose quartiles fell in the −1.75 to 1.0 range. Examiners agreed with photographic grading within ± 1 in more than 80% of gradings for vascularity and bleb height, within ± 1 in more than 75% of gradings for bleb wall thickness, within ± 2 in 61% of bleb width assessments, and ± 2 in 59% of diffuse component.
This bleb grading system is reproducible clinically and photographically. High levels of agreement between scores for photographs versus slit lamp examination were found for most categories, with good interobserver agreement for both photograph and slit lamp grading. Further refinement of scoring and reference photographs is required for optimization, especially for grading of bleb morphology.
From the *Wound Healing Research Unit, Institute of Ophthalmology and Moorfields Eye Hospital, London; †Ophthalmology Unit, Wellington School of Medicine, Newtown, New Zealand; ‡University of California, San Diego, La Jolla, California; §22 Harley Street, London; ¶Biostatistics, Research and Development Department, Moorfields Eye Hospital, London; and ‖Institute of Ophthalmology and Moorfields Eye Hospital, London.
Received for publication September 2003; accepted February 2004.
This work was supported by grants from the International Glaucoma Association, Moorfields Trustees, Capital Vision Research, The Healing Fund, Wellcome Grant 056045, and MRC Grant G9330070.
Reprints: A. P. Wells, Wound Healing Research Unit, Institute of Ophthalmology and Moorfields Eye Hospital, 11-43 Bath Street, London, EC1V 9EL (e-mail: email@example.com).