Optic Nerve Structure in Healthy Subjects—Reply

In reply

In a recent study,¹ we investigated structural and functional differences between the eyes of healthy blacks and whites, as blacks tend to be disproportionately affected by primary open-angle glaucoma.² We found that blacks had a significantly thicker RNFL in the superior and inferior quadrants and larger optic discs than whites. Our interpretation of these findings is that large discs may contain more nerve fibers than small discs. This is consistent with histological evidence showing that healthy blacks have larger optic discs than whites³ and that in monkeys, the number of nerve fibers is positively correlated with the size of the optic nerve.⁴ Savini and colleagues, however, have suggested that our results could instead be explained by an artifact associated with the fixed diameter used for the circular scans of the OCT. These scans are centered on the optic disc and performed at a fixed diameter of 3.4 mm. The OCT circular scans are therefore performed closer to the optic disc margin in large discs as compared with small discs. Previous studies^{5 - 7} have shown that in healthy eyes, the RNFL is thickest near the optic disc margin and becomes gradually thinner as the distance from the optic disc margin increases. Given that the OCT scans were taken closer to the optic disc margin in blacks (with larger discs) than in whites (with smaller discs) and that the RNFL is thicker near the optic disc margin, Savini and colleagues suggest that our results may be an artifact of the fixed-diameter scans. While this explanation is intuitive, we argue against it and welcome the opportunity to discuss our rationale.

We suggest that when interpreting RNFL thickness measurements, consideration should also be given to the total volume occupied by the nerve fibers. For example, if 2 optic nerves, 1 large and 1 small, contain the same number of nerve fibers, we argue that thicker RNFL measurements would be obtained at the optic disc margin of the small disc as compared with the large disc because it has a smaller circumference. The large disc, with its larger circumference, provides more space for the fibers at the disc margin, thus producing thinner RNFL measurements. While the RNFL mean thickness measurements differ, the volume occupied by the nerve fibers is the same in both optic discs. Therefore, 2 discs of different size but containing the same number of nerve fibers will have different RNFL mean thicknesses if measured at the optic disc margin. A valid comparison of RNFL mean thickness measurements in eyes with different disc sizes thus requires using a fixed diameter for circular scans. At any fixed distance from the center of the optic disc, RNFL mean thickness measurements should be the same for small and large optic discs containing the same number of fibers.

Carpineto et al⁸ compared RNFL mean thickness measurements obtained with the fixed-diameter (3.4-mm) and nerve head circle (0.85 mm from the optic disc margin) circular OCT scans. Their results show an increase in RNFL mean thickness with increasing disc size with the fixed-diameter scans, whereas the RNFL mean thickness was similar for all disc sizes with the nerve head circle scans (see Figure 1 for a schematic representation of their findings). They concluded that the relationship between optic disc size and RNFL mean thickness obtained with the fixed-diameter scans is an artifact, as the relationship disappears when using the nerve head circle scans. We believe, however, that their results can also be explained within our theoretical framework, which is based on the assumption that larger discs contain more nerve fibers. The results they obtained with the fixed-diameter scans are consistent with our view in that increasingly thicker RNFL measurements should be obtained with increasing disc size. However, if Carpineto and colleagues are correct and there is no true correlation between disc size and the number of nerve fibers, then measuring RNFL mean thickness with the nerve head circle scans should yield the pattern of results indicated by the solid line in Figure 2. At the same distance from the optic disc margin, thinner RNFL measurements should be obtained for the larger discs compared with the smaller discs because the larger circumference of the large disc provides more space for the fibers. However, Carpineto and colleagues failed to observe this pattern of results when measuring with the nerve head circle scans. Instead, their results are consistent with large discs containing more nerve fibers (thicker mean RNFL measurements) and smaller discs containing fewer fibers (thinner mean RNFL measurements) as indicated by the arrows and the dotted line in Figure 2.

In our study,¹ although racial differences in RNFL mean thickness did not reach statistical significance, a trend suggesting that blacks have thicker RNFLs than whites was observed. Our P value (.09) was relatively close to reaching statistical significance, and it is likely that this trend will become significant when a larger sample is tested. While our statistical power allowed us to detect differences of 8 μm, we observed a difference of 6.36 μm between the 2 groups. The results did reach statistical significance in the superior and inferior quadrants, which is in line with the trend observed for the RNFL mean thickness. In our article,¹ we mentioned that contrary to our findings, previous studies^{9 - 10} have reported thinner RNFLs in blacks. These studies used scanning laser polarimetry, and although the recent software version uses fixed-diameter scans, the earlier version used in these studies scanned at a distance relative to the optic disc margin. It is therefore not surprising that thinner RNFLs were observed for blacks (with larger discs) than whites (with smaller discs).

In summary, we argue that the combination of thicker mean RNFLs and larger optic discs that we observed in blacks in our study¹ is likely a reflection of the greater number of nerve fibers in larger discs. We do not attribute this finding to an artifact of the OCT associated with the fixed-diameter circular scans. However, it is possible that other factors not yet identified could influence the relationship between disc size and RNFL thickness as measured by OCT. Future studies with larger samples will shed some light on these issues.

Correspondence: Dr Sample, Department of Ophthalmology, University of California, San Diego, 9500 Gilman Dr, La Jolla, CA 92093-0946 (psample@glaucoma.ucsd.edu).

Financial Disclosure: None reported.