This study documents that macular fundus autofluorescence increases in a large proportion (76.8%) of patients with CSME. Moreover, retinal sensitivity decreases over areas with increased FAF, indicating that the function of the neurosensory retina deteriorates when autofluorescence increases.
DME is currently defined as increased retinal thickness. The functional impact of DME is usually determined by the quantification of BCVA.
9,31 But, BCVA does not depend only on macular thickness. The Diabetic Retinopathy Clinical Research Network reported that for a given degree of diabetic maculopathy, a wide range of visual acuity may be observed. Therefore, changes in retinal thickness have low predictive value compared with BCVA changes.
9
Microperimetry has recently been introduced in clinical practice for functional evaluation of DME.
10 –16 Microperimetry quantifies macular sensitivity, exactly correlating it to fundus characteristics, and determines retinal fixation characteristics. A significant inverse correlation between macular thickness and macular sensitivity has already been reported in many studies.
12 –15 More contrasting data exist about fixation characteristics in DME eyes
10,11,14 —mostly due to the differences in examined populations, especially differences in DME duration.
10,11,14 Vujosevic et al.
10 reported in a well-defined group of CSME eyes that location and stability of fixation were normal, except when subfoveal hard exudates were present. Microperimetry may add some new information in the management of DME that is complementary to OCT and visual acuity data.
12 Functional evaluation of DME is important for two reasons: first, morphologic evaluation is useful to assess the severity of DME and treatment outcomes, but it must correlate to visual function, to give better DME prognostication. Second, better correlation of structural and functional parameters may enable identification of new DME patterns, which may be responsible for different responses to local treatments. Therefore, treatments based on OCT macular thickness or FA leakage alone may be not as efficient as expected, giving rise to different outcomes. This fact may suggest the presence of different DME phenotypes. The determination of new patterns corresponding to these phenotypes may be useful in tailoring DME treatment.
Although classification of DME may vary when using different diagnostic methods and terms such as focal and diffuse DME are not yet standardized, several studies have recently shown strong correlations between OCT findings and FA patterns in DME.
17,32 –34 These correlations address the changes in intraretinal structure. Large foveal cysts, located in the outer nuclear layer and/or Henle's layer found on OCT correspond well to petaloid cystoid leakage patterns on FA.
32,34 With the advent of new high-resolution OCT, fine intraretinal structure changes caused by cystic spaces have been visualized and could be better compared to histopathologic reports.
35,36 In the present study, a high proportion of agreement in the evaluation of cystoid edema with OCT and FA was found, although a time-domain machine was used. On the contrary, OCT and FA disagree in detecting subfoveal neuroretinal detachment. Whereas SNRD is easily detected on OCT, it does not have any specific pattern on FA.
FAF has been used mostly for structural evaluation of age-related macular degeneration, inherited macular dystrophies, and cystoid macular edema of different origins, but FAF characteristics in DME are still poorly understood.
18 –22,37 –39 Bessho et al.
37 documented increased FAF in all examined eyes with cystoid macular edema of different origin corresponding to the petaloid shape on FA and OCT cysts. Pece et al.
38 described three different increased patterns of FAF (multicystic increased, single cyst increased, and combined single- and multicystic increased FAF) in patients with cystoid DME that correlated positively with FA and OCT findings. In the present study, FAF parameters correlated to both structural and functional parameters more commonly used in DME. The presence of iFAF was associated with functional and structural impairment of the macula. The iFAF pattern group had poorer macular sensitivity than did the normal FAF group. These data indicate that DME with iFAF pattern is, at least functionally, more severe than DME with a normal FAF pattern.
Other studies involved analysis of the correlation between FAF imaging and microperimetry in age-related macular degeneration, and the results are comparable to those obtained in this study.
40,41 In particular, there was a significant correlation between presence of iFAF areas and decreased retinal sensitivity, regardless of visual acuity.
40
In the present study, an increased FAF pattern (both single- and multiple-spot iFAF) was associated with cystoid edema pattern on OCT in approximately 89% of cases. Although the difference between single and multiple iFAF may sometimes be subtle, adequate clinical training overcomes this limitation.
42 McBain et al.
22 found high sensitivity and specificity of FAF examination in detecting cystoid edema, proposing FAF detection as a valid and noninvasive alternative to FA in evaluating cystoid macular edema of different origins.
The origin and significance of FAF in DME is still unclear. FAF is thought to visualize the distribution of lipofuscin in the RPE.
20,43 –45 In the latter, lipofuscin is mainly produced by incomplete degradation of photoreceptor outer segments, and it accumulates with age.
43
In diabetic retinopathy, age-related accumulation of lipofuscin in the RPE, due to photoreceptor outer segment phagocytosis, does not appear to be a relevant pathogenic mechanism. But, lipofuscin contains a large number of molecules that are mainly peroxidation products of proteins and lipids.
20 Thus, lipofuscin seems to be also an indicator of oxidative damage within the retina. A recent histologic study found that lipofuscin accumulates in microglia much more than in the RPE.
46 This hypothesis could partly explain our results. It is well-known that microglia are activated by diabetes; this activation could determine the oxidation of proteins and lipids that accumulate in the microglia due to ongoing diabetes.
46 Thus, we hypothesize that the iFAF areas observed in DME are caused by accumulation of oxidative product induced by activated microglia.
46
Another hypothesis is mainly related to the mechanical effect of cystoid macular edema: Cysts are mostly located in the outer plexiform and inner nuclear layers, where there is a maximum accumulation of luteal pigment.
22 Cysts may displace luteal pigment preventing the normal blockage of foveal FAF signal at the level of each of them.
22,45 Moreover, Bessho et al.
37 documented increased FAF by 488-nm excitation in all examined eyes with cystoid macular edema, but not by 580-nm excitation, concluding that increased FAF is a pseudoautofluorescence due to a window defect, as in FA. As a consequence of this hypothesis, when cysts are reabsorbed, foveal FAF signal should disappear. This conclusion is in contrast with our experience in the follow-up of diabetic patients treated with laser photocoagulation in whom cysts reabsorbed but iFAF remained (Vujosevic S, et al.
IOVS 2009;50:ARVO E-Abstract 1377). Although this phenomenon may be explained by residual displaced luteal pigment, it is also reasonable that it is related to persisting activated microglia.
However, more studies are needed to determine the exact origin of increased FAF signal in diabetes. FAF imaging could assist in identifying different DME patterns to improve macular edema classification and treatment in diabetic patients.
The authors thank Fabiano Cavarzeran, ScD, for statistical analysis of the data.