Article Text

Download PDFPDF

Retinal nerve fibre layer loss in diabetes mellitus without retinopathy
  1. B Skarf
  1. Department of Ophthalmology, Henry Ford Health Sciences Center, Detroit, Michigan, USA; bskarf1{at}

    Statistics from

    New study provides no new information

    Many diabetics worldwide might benefit if a simple, reliable, objective, and non-invasive technical method were available to detect and monitor the development of retinopathy and visual deficits without requiring a professional examination. With this ultimate goal in mind, Lopes de Faria and colleagues describe their initial use of scanning laser polarimetry (SLP) to measure retinal nerve fibre layer thickness in diabetics without retinopathy in this issue of the BJO (p 725).

    Evaluation of the retinal nerve fibre layer (RNFL) as a means of assessing optic nerve health has been a clinical and investigational tool for at least 30 years. Ophthalmoscopic scrutiny of the nerve fibre layer was first advocated by Hoyt1–3 and this approach was soon enhanced by the description of photographic manipulations designed to increase the contrast between retinal areas with different nerve fibre layer thicknesses.4 For most ophthalmologists, these qualitative methods were difficult to apply and had limited usefulness in monitoring progression of disease in a clinical setting. Furthermore, they depended greatly on the clarity of the ocular media. Now, with instruments capable of objective determination of nerve fibre layer thickness, substantial interest has focused on the significance of early thinning of the nerve fibre layer in glaucoma, papilloedema, and a variety of other optic neuropathies.

    Lopes de Faria and co-workers used the GDx nerve fibre analyser (Laser Diagnostic Technologies, San Diego, USA) to provide the first quantitative assessment of nerve fibre layer thickness in diabetics. Their small sample of 12 patients, between the ages of 18 and 40, have had type 1 diabetes mellitus for a minimum of 10 years, with no retinopathy detectable by clinical examination or fluorescein angiography (level 1, modified Airlie House classification). In comparing this sample with an age and sex matched control group without diabetes, the authors demonstrated that the diabetic group has statistically significant thinning of the nerve fibre layer in the quadrant superior to the optic disc. There is no statistical difference between control and diabetic group data for nasal, temporal, or inferior retinal quadrants. The authors offer no explanation for this selective topographic difference in RNFL thinning.

    The usefulness of SLP in the early detection of RNFL thinning in diabetes should not be inflated

    Although interesting because of the use of an objective technique, this study provides no new information. It confirms what we have known for 8 years from the only other published studies that describe RNFL findings in diabetics without retinopathy5,6 (appropriately cited by our authors).

    The significance of this finding and the usefulness of SLP in the early detection of RNFL thinning in diabetes should not be inflated. Firstly, the sample chosen for this study is very small and though the statistical analysis shows significance at the p=0.03 level, the statistical approach used in this study is suboptimal. Given the multiple comparisons that were made between different sets of data, the probability that one of the measures would have been falsely positive was increased by the approach used. A multivariate analysis would have been more appropriate. It is also worth noting that for two pairs of data sets (inferior and nasal polar integrals) the SD of the data for one group (that is, patient or control) is an order of magnitude larger than the SD of the data from the other group. Thus, data for inferior and nasal polar integral measures could not possibly achieve statistical significance. Furthermore, review of the data presented in Figure 2 shows that all except two or three of the patients studied fall within 2.5 SD of the mean of the control group. Thus, these data will not allow the authors (or us) to determine whether each individual specific measure of superior RNFL thickness in most of their diabetic patients is abnormal.

    In conclusion, this study, supporting earlier qualitative observation, hints at the possibility that careful RNFL thickness measurements might help detect early changes in diabetics. We are still no closer to recognising these measures as being clinically useful and it would be premature to begin evaluating the RNFL of individual diabetic patients using these techniques. Properly designed, prospective, longitudinal clinical studies on larger populations of patients without retinopathy are mandated before we can accept and recommend the diagnostic and predictive value of RNFL measures made by this or any other methodology.

    Note in Proof

    New study provides no new information


    View Abstract

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles