Peripapillary retinal blood flow in normal tension glaucoma

Aims—To determine if normal tension glaucoma (NTG) patients diVer from age matched controls in blood flow to the peripapillary retina, as measured with confocal scanning laser Doppler flowmetry (cSLDF; “Heidelberg retinal flowmetry”). Methods—12 NTG patients and 12 age matched controls were compared using (a) 10 × 10 pixel boxes (the instrument default sample size), taken from the nasal and temporal peripapillary retina, (b) the average from two of these boxes, and (c) every qualifying pixel within the peripapillary retina. Results—Patients and controls did not diVer in blood flow measured using the default sample from a single 10 × 10 pixel box, placed in either the temporal or nasal peripapillary retina, or expressed as the average from these two boxes. However, in histograms using every pixel from the peripapillary retina, NTG patients displayed significantly higher percentages of minimal flow pixels (defined as less than one arbitrary unit of flow: 30% v 19%, p <0.01), and significantly lower flow in the 25th, 50th, and 75th percentile flow pixel (each p <0.05) than did age matched controls. Conclusion—NTG is characterised by reduced blood flow in the peripapillary retina, a result suggesting that blood flow deficits accompany, and perhaps may contribute to, disease development in these patients. (Br J Ophthalmol 1999;83:466–469) Glaucoma may result from the programmed death of retinal ganglion cells, 2 as initiated by complex factors that include mechanical compression and ischaemia. Seen in these terms, it is apparent that defining the level of blood flow in and around the optic nerve head is an important goal for visual science—to test hypotheses regarding blood flow and disease, to determine disease severity, and to monitor treatment interventions. Several lines of evidence suggest that patients with normal tension glaucoma (NTG) may suVer from ocular blood flow deficits. In these people, vascular resistance downstream from the central retinal and posterior ciliary arteries is increased, 5 choroidal filling times are prolonged, areas of indocyanine green hypofluorescence are increased in the peripapillary region, and diVuse ischaemia may exist throughout the brain. However, these haemodynamic markers provide no direct information about perfusion of the retina or optic nerve head. In this study, confocal scanning laser Doppler flowmetry (cSLDF) was used to determine capillary blood flow in the peripapillary retina near the optic nerve head. Previous studies that have used this technique have failed to detect a perfusion diVerence between controls and patients with NTG. 11 However, those experiments utilised the default 10 × 10 pixel sampling box, a method that exhibits a high coeYcient of variation when repeated on a weekly basis. In this study, in addition to conventional analysis, we utilised a new methodology that includes every qualifying pixel within the entire cSLDF image. Besides measuring the distribution of low and high flow pixels within the sample, this novel method reduces the coeYcient of variation of repeated measurements by nearly 50%. Patients and methods