4.4.2.1 Normal group
As per the inclusion criteria; a normal subject did not have any ocular disease that could affect their field of vision directly or indirectly through medications or other influences. In the intra- group analysis, when the inter-sector/hemi-ring differences between corresponding hemifields were tested for significance, a comparison between the 6 sectors and 5 hemi-rings on each hemifield was carried out. Only one pair of corresponding sectors was statistically signifcant (SS5 -
129
IS5). This result was concordant with the hypothesis but also logical thinking, as normal subjects should not have any signifcant difference between the two hemifields. Only one sector yielded a statistically signifcant difference between the two hemifields, which could be attributed to the wide normal range of normal SNR responses among the normal group. Normal SNR response in this study showed a mean of 2.85. However the variability was the highest amongst the 3 groups (SD = 0.503), which could be expalined in the light of normal distribution of the data, as a normal variation. Many studies have shown evidence of a wide range of SNRs among normal individuals
(39,71,131-134)
. The variability in the normal subjects showed a mean SNR varied by a factor of 3, from about 2.0–6.0 in many studies (132-136). The relatively small SNRs in a few individuals can be traced to high noise levels. Most of these variations were attributed to a variation in the size of the signals. It is this range of SNR values that creates a problem for tests of significance, especially those based upon analyses of monocular mfVEP test. Variability issues even among normal individuals pushed researchers into one of two directions to overcome this problem and make their data valid. (136). Moreover, in similar studies of monocular mfVEP testing the SNR values obtained from individual segments showed 7 or even 10 folds (SNR average 9.5) which is close to the values obtained and recorded for normal subjects in this study, which ranged between (2.78 and 13.8) throughout all 58 segments, and of course following the cotrical representation rule of higher values more centrally and lower towards the periphery. A difference in this study is that averaging of the SNR values was used from a specific group of segments representing a sector or a hemi-ring, and in this case the SNR value is less and more consistent all over the entire field with very minimal variation between peripheral and central sectors and hemi-rings. Each pair of sectors and hemi-rings had paired t-test performed to check for statistical significance. The major problem with running a mfVEP monocular test is the need to reference values to establish a comparison, which is why it was stated that the variability of mfVEP among normal individuals necessitates the need for establishing a normative database for the test with accepted range of standard deviation from the mean values for central and peripheral responses of the visual field. The variability of mfVEP among normal individuals made it necessary for researchers to standardize a suitable normative database that could be used for patients with an accepted range of standard deviation from the mean values for central and peripheral responses of the visual field. However, this normative data cannot be standardized as a universal database in any testing equipment, but many commercially available software packages use some form of normative database in their testing procedure to produce grading scale with probability maps; VERIS™ and AccuMap® systems are examples of these software. The most suitable method suggested by many authors (19,39,159,160) is for each clinic to create its own age-related normative data. The advantage of the hemifield
130
sector analysis protocol is that it provides an intrinsic method of standardization and referencing of SNR values by comparing within the same eye corresponding sectors and hemi-rings, which does not require any external normative data in identifying the difference between normal and glaucomatous focal visual field defects. In addition, what determines the usefulness of a test that measures the progression in glaucomatous visual field defects is its good repeatability. mfVEP has shown good repeatability in many studies, even better than SAP protocols (142,145,146,161,162).
4.4.2.2 Glaucoma suspect group
Glaucoma suspects present uncertainty and confusion to many clinicians when reaching a diagnosis, especially as the main tool for visual assessment is the SAP visual field testing. This group of patients usually presents with one or more positive clinical findings or risk factors for glaucoma, but without definite optic neuropathy or established glaucomatous visual field loss. In this study group, there was a lack of homogeneity in the clinical findings and level of expected defects among patients. Some patients had repeated abnormal visual fields with focal defects that do not match a healthy disc appearance, but high intraocular pressure was reported. In other patients they were considered glaucoma suspects on the basis of a suspicious optic disc appearance, disparity in cup-to-disc ratio between the eyes, persistently high intraocular pressure, and positive family history, but in the absence of significant visual field defects. The lack of homogeneity can be attributed to the nature of the diagnosis, and how it is related to many variable findings and clinical signs that each one of them can be considered a reason to suspect glaucoma. However, the presence of significant visual field defects could have moved those patients into the category of glaucoma patients, since persistent and reproducible visual field defects in the absence of other findings could be viewed as an early functional deficit. If the visual field assessment is not accurate or equivocal then the diagnosis of glaucoma could be missed or the treatment may commence without conclusive evidence that a glaucomatous process is ongoing. Using the sector hemifield analysis protocol, it was found that 4/6 sectors showed a statistically signifcant SNR difference when compared to their corresponding fellows, while 1/5 hemi-rings was statistically significant. This result is expected based on the nature of diagnosis in this group, as most of the SNR values were very close to normal range, but distinctly far from the glaucoma SNR range of values. The presence of this signifcant difference between the two opposite hemifields could give an indication that there are focal depressions in the responses in certain parts of the visual field, and the mfVEP was able to detect the focal depressions and their location. Most of the glaucoma suspect patients enrolled in this study presented with high
131
intraocular pressure as a major risk factor for glaucoma. These results are not conclusive, nor can they confirm that the patients found to have a statistically significant difference between the responses in the two opposite hemifields, have an early process of glaucoma taking place. There is a lack of longitudinal data in this research to establish whether these patients will develop glaucoma in the future. This limitation necessitates the need for longitudinal study. However, numerous studies have reported that mfVEP can detect early glaucomatous field defects by a significant reduction in the SNR response, or ampitude values. Recent evidence suggests that mfVEP can have clear role in monitoring and detecting progression of glaucoma based on good repeatability figures (19,145,159,170,171). There were many examples where HFA and mfVEP were not consistent and did not agree in the detection of visual field defects. These results are in agreement with Chen (145), Goldberg (134) and Greenstein (172) who showed that mfVEP detected damage missed by HFA. Furthermore, abnormal mfVEPs have been reported in patients with normal visual fields assessed with the HFA. These findings have been attributed to the subjective responses of the patient in SAP and the learning curve that can mask shallow and early defects. Obviously, there is no learning curve required for mfVEP test, which can make the detection of small visual field changes more reliable. However, it is important to emphasize that in order to perform a mfVEP test accurately, both the operator and interpreter of the test should have a minimum level of technical skills and sufficient knowledge of the test process. This requirement can be considered as a form of “learning curve”, which can influence the reliability and accuracy of the test, similar to that of the SAP test.
4.4.2.3 Glaucoma group
The mfVEP was extensively studied in the assessment of glaucomatous visual field defects and was found to have excellent results, yielding high sensitivity and specificity (39,73,120,123,174,176,225). In this study, the majority of patients suffered from chronic open angle glaucoma, which is the most common type of glaucoma worldwide (6,7). Recruitment was designed to give a representative sample of patients with variable degrees of severity represented by their visual field defects. This sample presented with defects ranging between early focal scotomas to the deep and extensive visual field losses which are a feature of advanced glaucoma. Most of the patients (13/20) had asymmetrical glaucomatous field changes between eyes, the rest (7/20) had symmetrical mild to moderate field changes defined by the Hodapp – Anderson classification (216,219) . The results show that the majority of patients had asymmetrical glaucomatous visual field defects between the two eyes, which is a good confirmation of how important the monocular mfVEP test analysis is. The
132
monocular analysis is more beneficial in these patients simply because for inter-ocular analysis, it is assumed that one eye is a “normal” reference for the glaucomatous one and therefore it will not detect the defects accurately. The presence of symmetrical defects usually influences the ability of inter-ocular tests to idenfity defects. This has been confirmed by many studies, thus the inter- ocular comparison test might not detect bilateral damage located in corresponding field locations
(39,137)
. The mean SNR value for glaucoma group was 1.70 ± 0.412, which is low average response, close to poor (SNR = 1). This SNR average confirms the ability of the HSA protocol to identify glaucomatous visual field defects accurately. Only 8/36 eyes exhibited advanced glaucomatous changes with symmetrical severe field loss across the midline in both hemifields. The remainder of the sample (28/36) showed clear differences between the two corresponding hemifields. In such patients, the hemifield sector analysis protocol used in this research will easily detect isolated focal defects in one hemifield sector or hemi-ring by direct comparison to their fellows in the opposite hemifield. The results confirmed this hypothesis, and it was found by running the inter- sector/hemi-ring differences between corresponding hemifields in 6 pairs of sectors and 5 hemi- rings on each hemifield that all sectors and hemi-rings were statistically signifcant when compared to their coresponding fellows in the opposite hemifield. This means that regardless of the level or severity of the glaucomatous visual field defect, the hemifield sector analysis protocol of mfVEP was able to identify the defect and its location and highlight a low response due to the field defect. These results also in agreement of several studies carried out over the past two decades, where the ability of mfVEP test to localize and identify already established glaucomatous visual field defects was confirmed (39,125,133,142,144,146,174,178,225).