Chapter 214 – Visual Field Perimetry in Glaucoma
ELLIOT B. WERNER
• The visual field is the space that one eye can see while remaining fixed.
• Measurement of the extent of the visual field by projecting targets onto a curved surface is called perimetry.
• Modern quantitative perimetry measures the differential light threshold at various locations in the visual field.
• Glaucoma produces characteristic but not pathognomonic changes in the visual field.
• At present, perimetry is among the best tests to determine the extent of glaucomatous damage to visual function and whether or not visual loss is progressive.
Glaucoma, of course, is not a single disease but a collection of conditions that have in common the tendency to produce a characteristic optic neuropathy characterized mainly by cupping. Whereas some glaucomas are acute and associated with symptoms at onset, most glaucoma patients have a chronic, slowly developing disease that does not produce prominent symptoms until optic nerve damage and visual loss are far advanced. Glaucoma is treated in order to preserve vision. Tests of visual function, thus, are of critical importance in evaluating the glaucoma patient and in guiding treatment. Despite recent advances in optic nerve and retinal nerve fiber layer evaluation and despite many recent studies of a variety of visual functions in glaucoma, white light perimetry remains the most reliable widely used tool to determine the likelihood of a patient suffering significant functional impairment as a result of glaucoma.
As a pure diagnostic tool, perimetry has a number of shortcomings.  Other modalities may be more useful for population screening. Automated perimetry, however, compares favorably with other clinical techniques for detecting glaucomatous damage. As an indication of the extent of glaucomatous damage and as an aid in determining whether an individual with glaucoma is deteriorating, perimetry is indispensable.    
The existence of the field of vision was known to the ancients, and Hippocrates is said to have described hemianopias. The French physicist Mariotte described the physiological blind spot in the 17th century. Young and Purkinje described and measured the limits of the visual field in the early 19th century, and von Graefe was the first to use measurement of the visual field clinically in the 1850s.
The modern tangent screen was introduced by Bjerrum in 1889. He used this device not only to measure the peripheral extent of the visual field but also to detect localized defects or scotomas in the central visual field. The typical arcuate scotomas seen in glaucoma still bear Bjerrum’s name. Despite efforts to improve on Bjerrum’s original design, such as the Autoplot, which projects a spot of light onto a wall-mounted screen, the black tangent screen remains the standard tool for campimetry, that is, the measurement of the visual field on a flat surface.
Perimetry refers to the measurement of the visual field on a curved surface and has largely replaced campimetry in modern clinical practice. The first perimeters were arc perimeters that, like the tangent screen, used small round objects as test targets. Light projection arc perimeters, such as the Aimark, were introduced in the 1930s, and the development of the Goldmann hemispheric projection perimeter in 1945 ushered in the modern era of quantitative perimetry.
Computer technology was combined with visual field testing in the mid-1970s, resulting in the introduction of the first automated perimeters, the television campimeter of Lynn and Tate, the Octopus device of Fankhauser, and the Competer of Heijl and Krakau.
During the ensuing decade a number of automated perimeters were introduced, and many are no longer manufactured. There are now several automated visual field testing devices on the market, but the two most widely used systems are the Octopus perimeter marketed by the Swiss firm Interzeag and the Humphrey Visual Field Analyzer marketed by the American firm Humphrey Instruments. Automated perimetry has largely replaced manual perimetry in clinical practice because of its superiority in detecting glaucomatous visual field loss.
Several detailed reviews of the history of visual field testing have been published.  
PURPOSE OF THE TEST
There are two steps in diagnosing glaucomatous visual field loss using automated perimetry. The first is to determine whether the visual field is normal. If the visual field is abnormal, the second step is to decide whether the visual field abnormality is due to glaucoma or something else. The second step is actually the easier of the two. Differentiating the normal from the abnormal field is not straightforward.  It requires a knowledge of the range of visual field responses in the normal population, an understanding of probability, and the ability to interpret detailed statistical analysis of the visual field data. The essential purpose of visual field testing in glaucoma is to determine the extent of functional visual loss and whether or not it is progressive.
Use of Probability Statistics to Define the Normal Visual Field
When applied to perimetry, the term normal describes the range of test results found in the nondiseased population. The range of normal has been determined experimentally, and the results
are stored in the computer memory of most automated perimeters. This allows the comparison of an individual patient’s visual field with the expected normal values.  
Because of the wide range of normal, one cannot say with certainty that a particular visual field is normal or abnormal. One can, however, determine the likelihood of finding a particular visual field result in a normal individual. If that likelihood is very small, the visual field is probably abnormal. Specific examples of visual fields in this chapter are taken from the Humphrey Visual Field Analyzer. Similar results and analytic software are available with the Octopus and other automated perimeters.
The determination that a visual field is within the normal range cannot be made by simple inspection. Statistical analysis is necessary. Statistical software packages available with some perimeters provide probability statements about the visual field data that allow the clinician to determine the likelihood that a visual field is normal.   If all statistical parameters are within the normal range, chances are that the visual field is normal. The sensitivity of automated threshold perimetry for detecting visual field defects is very high. It is extremely unlikely that a patient with a clinically significant visual field defect would have a normal result. The opposite, however, is not always true.
Many otherwise normal patients have a visual field that may be abnormal because of the large number of artifacts that can occur during automated visual field testing. In other words, the
Figure 214-1 Printout of a normal visual field from the Humphrey Visual Field Analyzer. The reliability indices, fixation losses, false-positive errors, and false-negative errors are found in the upper left-hand corner a few lines below the patient’s name. The Glaucoma Hemifield Test result is found just below the gray scale interpolation and is “within normal limits.” The global indices are found in the lower right-hand corner just below the Glaucoma Hemifield Test. Neither the MD nor the PSD is flagged with a P value, indicating that both are within the expected range of normal. The numerical and graphic representations of the total and pattern deviations are found where indicated in the left and center lower portion of the printout. There are no probability symbols in either the total or pattern deviation, indicating that the threshold value of each test location is within the normal range for the patient’s age.
specificity of automated perimetry is often not as high as clinicians would like. When performing perimetry on patients suspected of having glaucoma, it is important to distinguish the visual field that appears abnormal because of artifact from the visual field that is truly abnormal as a result of glaucoma or some other disease such as cataract, retinal disease, or neurological lesions. Statistical analysis must be combined with other clinical data, experience, and the ability to recognize specific patterns of visual field loss related to specific diseases or artifacts.    
The reliability indices are found in the upper left-hand corner of the printout of the Humphrey Visual Field Analyzer ( Fig. 214-1 ). Reliability is evaluated by measuring fixation losses, false-positive and false-negative responses. The fixation loss rate measures how often the patient fails to fixate the central target. In the (HFA I), the fixation loss rate relates to the number of times a patient responds to a target placed in the blind spot. In the newer model (HFA II), fixation is monitored by an eye tracker. A real-time display of eye movements during the test is presented across the bottom of the printout ( Fig. 214-2 ).
The false-positive error rate refers to the number of times a patient responds when no test target is presented. The false-negative
Figure 214-2 Visual field showing an elevated fixation loss rate of 10/15. The eye tracker tracing across the bottom of the printout shows larger eye movements during the first half of the test. Eye movements decreased and fixation presumably improved during the second half of the test. False-positive and false-negative errors are also slightly elevated. The presence of an elevated reliability index does not necessarily mean that the results of the test are not valid. Repeatability and correlation with other clinical findings are more important criteria. This visual field shows a well-defined superior nerve fiber bundle defect that was repeatable and correlated well with the appearance of the optic nerve, indicating that the result is believable despite the poor fixation.
error rate refers to the number of times a patient fails to respond to a suprathreshold (very bright) target placed in a seeing area of the visual field   ( Fig. 214-2 ). The standard full-threshold test algorithm measures false positives by presenting an audible clue when no test target is displayed. False negatives are measured by presenting suprathreshold targets in seeing areas. The Swedish Interactive Thresholding Algorithm (SITA) on the HFA II calculates false-positive and false-negative rates from the time between the presentation of the target and the patient’s response during threshold determination. 
The reliability indices are an indicator of the extent to which a particular patient’s results may be reliably compared with the normal range of values stored in the computer memory. Automated perimetry in patients with poor reliability has lower specificity and sensitivity for the detection of visual field defects. Test results must be interpreted with caution in these patients,    although useful results can often be obtained despite what appears to be poor patient performance ( Fig. 214-2 ).
A high false-positive rate is often associated with the patient who responds frequently without regard to whether a target is seen, the so-called “trigger happy” patient. This may result in a visual field with abnormally high decibel thresholds, which is not interpretable ( Fig. 214-3 ).   High false-positive or false-negative response rates are associated with alterations in the overall sensitivity of the visual field that could make detecting defects more difficult. High fixation loss rates due to eye movements have been associated with increased variability of the visual field responses and increased difficulty in detecting scotomas.  
Figure 214-3 Visual field of a “trigger happy” patient who frequently responds at times when no test target is presented. The result is a high false-positive error rate, an unphysiologically elevated mean deviation (MD), and an abnormally high sensitivity, which is seen as white areas in the gray scale interpolation, the so-called white field artifact. The Glaucoma Hemifield Test shows “abnormally high sensitivity.” There are many probability symbols in the pattern deviation not seen in the total deviation.
The global indices are found in the lower right-hand corner of the Humphrey printout (see Fig. 214-1 ). The mean deviation (MD) is a measure of the average difference between the threshold value of each test location and the age-corrected normal value. The pattern standard deviation (PSD) is the standard deviation of the mean difference between the threshold value at each test location and the expected normal value. It is a measure of the extent to which the threshold determinations at different locations in the visual field differ from each other.  The loss variance (LV) of the Octopus system, although calculated differently, provides similar information.
The calculation of the global indices is weighted to give greater importance to the test locations near fixation and less importance to more peripheral locations.  The formulas are fairly complex and beyond the scope of this chapter. Interested readers are referred to Anderson and Patella.
If a global index is outside the expected normal range, a P value will appear next to it ( Fig. 214-4 ). The P value represents the proportion of normal subjects in which an index of that value is found. For example, if P <1% appears next to the MD, fewer than 1% of normal subjects of that age have an MD at that level. Any global index with a P value less than 5% has a high probability of being abnormal.
The MD is mainly an index of the size of a visual field defect. It is not, as is widely believed, an indicator of generalized depression of the visual field, but the MD is very sensitive to generalized loss of sensitivity. Purely localized defects that are large enough, however, also affect the MD.
Figure 214-4 Visual field of a patient with a nearly pure generalized loss due to cataract. The mean deviation (MD) is significantly below the expected range of normal at the P <0.5% level while the pattern standard deviation (PSD) is within the normal range. The total deviation shows many probability symbols, most of which are not found in the pattern deviation. The Glaucoma Hemifield Test shows “generalized reduction of sensitivity.”
TABLE 214-1 — INTERPRETATION OF THE GLOBAL INDICES ON THE HUMPHREY VISUAL FIELD ANALYZER*
Visual field probably normal ( Fig.214-1 )
Generalized loss of sensitivity ( Fig.214-4 )
Small localized defect ( Fig.214-5 )
Large defect with a significant localized component ( Figs.214-6 and 214-7 )
* Artifacts may cause the global indices to be abnormal in the absence of any pathological cause of visual field loss.
† The PSD is the equivalent of the loss variance (LV) of the Octopus system for purposes of interpretation.
TABLE 214-2 — INTERPRETATION OF THE TOTAL AND PATTERN DEVIATION ON THE HUMPHREY VISUAL FIELD ANALYZER
Probably normal ( Fig.214-1 )
Pure generalized loss ( Fig.214-4 )
Pure localized loss ( Figs.214-5 and 214-6 )
Mixed localized and generalized loss ( Fig.214-7 )
No or few symbols
Trigger happy patient ( Fig.214-3 )
The PSD is an index of localized nonuniformity of the surface of the hill of vision. It is sensitive to localized visual field defects and is not affected by purely generalized loss of sensitivity.
By looking at the MD and PSD, it is possible to anticipate the nature of any visual field defect before inspecting the rest of the visual field data. The interpretation of the global indices is summarized in Table 214-1 . If both the MD and PSD are abnormal, the patient may have either a mixed defect with both generalized and localized loss or a purely localized defect large enough to affect the MD. When both MD and PSD are abnormal, however, it is impossible to determine whether there is any significant generalized depression without inspecting the remainder of the visual field data.
Total and Pattern Deviation
The total and pattern deviations are found as arrays of numbers and graphic plots in the center and lower portions of the printout (see Fig. 214-1 ). The total deviation represents the difference between the measured threshold of each individual test location and the age-corrected normal value for that location. The actual measured thresholds are shown in the array of numbers in the upper central portion of the printout to the left of the gray scale.
The pattern deviation represents the difference between an adjusted threshold of each individual test location and the age-corrected normal value for that location. The pattern deviation is derived from the total deviation by adjusting the measured thresholds upward or downward by an amount that reflects any generalized change in the threshold of the least damaged portion of the visual field. The information in the total deviation, thus, may be thought of as a combination of generalized plus localized change. The information in the pattern deviation represents purely localized change. Although the interpretation of the pattern deviation and pattern standard deviation is similar and the similarity of the names may be confusing, they are not identical and should not be confused.
The graphic probability plots indicate how frequently a total or pattern deviation value at a particular test location will be found in the normal population. There are four symbols ranging
Figure 214-5 Visual field consistent with a small, localized defect. The MD is within the normal range while the PSD is outside the expected range of normal at the P <2% level. The total and pattern deviation plots show almost identical arrays of probability symbols. The Glaucoma Hemifield Test, which is very sensitive to small differences in threshold between the superior and inferior hemifields, shows “outside normal limits.”
from P <5% to P <0.5%. A black square, for example, indicates that the total or pattern deviation value for that test location will be found in fewer than 0.5% of normal subjects. Groupings of symbols in a portion of the visual field, therefore, indicate a high probability of an abnormality there.    The interpretation of the total and pattern deviation is summarized in Table 214-2 .
Glaucoma Hemifield Test
The Glaucoma Hemifield Test provides information about the difference between the superior and inferior halves of the visual field.  The Glaucoma Hemifield Test evaluates the differences in threshold of mirror image groups of points on either side of the horizontal midline. There are six interpretive messages that may appear, depending on the relationship of the thresholds in the superior and inferior halves of the field.
“Within normal limits” (see Fig. 214-1 ) means that there is no significant difference between the superior and inferior halves of the fields and the overall sensitivity is within the 99.5% range of normal.
“Outside normal limits” ( Fig. 214-5 ) appears when the threshold differences between the groups of points compared in the superior and inferior halves of the field are greater than would be expected in 99% of the normal population.
“Borderline”( Fig. 214-8 ) appears when the threshold differences are greater than would be expected in 97% of the normal population but not as great as in “outside normal limits.”
“General reduction of sensitivity” (see Fig. 214-4 ) appears when the overall sensitivity of the least damaged portion of
the visual field is depressed below the 99.5% range of normal but there is no significant difference between the superior and inferior halves of the field.
“Abnormally high sensitivity” (see Fig. 214-3 ) appears when the overall sensitivity is higher than expected in 99.5% of the normal population. This message is found most often in the presence of a high false-positive rate and usually represents an artifact of testing.
“Borderline” combined with “general reduction of sensitivity” appears in patients with a significant generalized loss of sensitivity and a residual moderate difference in the sensitivity of the superior and inferior hemifields.
Figure 214-6 Visual field consistent with a large defect with a significant localized component. Both the MD and the PSD are outside the expected range of normal at the P <0.5% level. Inspection of the global indices alone does not allow a determination of the relative amounts of generalized and localized loss. The number and distribution of probability symbols in the total and pattern deviation are very similar, indicating that this is almost a pure localized defect. The Glaucoma Hemifield Test is “outside normal limits,” reflecting the marked difference in sensitivity between the superior and inferior hemifields.
The specificity and sensitivity of the Glaucoma Hemifield Test for detecting nerve fiber bundle visual field defects are quite high, especially if consistently abnormal results are obtained after repeated testing. In the presence of the message “within normal limits,” it is very unlikely that a visual field defect of the type seen in glaucoma is present. On the other hand, although an abnormal Glaucoma Hemifield Test may be due to an artifact, the presence of a glaucomatous visual field defect is likely and should be carefully evaluated. 
UTILITY OF PERIMETRY
The management of glaucoma includes perimetry at regular intervals. All glaucoma patients capable of cooperating for the test should be tested at the time of initial diagnosis. Perimetry should
Figure 214-7 Visual field consistent with a large defect with a significant localized component. The MD and PSD are both outside the expected range of normal. The graphic representation of the total and pattern deviation shows many probability symbols in the total deviation. Many of the symbols in the total deviation are not seen in the pattern deviation. This is typical of a mixed generalized and localized defect (contrast Fig. 214-6 ). The Glaucoma Hemifield Test is “outside normal limits.” This type of field defect is often seen in glaucoma patients who also have a cataract or a small pupil. The localized component of the defect seen in the pattern deviation is consistent with an inferior and superior nerve fiber bundle defect.
then be repeated within a few weeks in order to have at least two baseline tests for comparison with subsequent tests. In patients with inconsistent results or with a significant learning effect, more than two baseline tests may be required. For ocular hypertensive patients or other patients with normal fields, perimetry every 12 to 18 months may be adequate. In patients with visual field loss, the frequency of testing will depend on the severity of the patient’s glaucoma and the risk for future progression. For such patients, two or more visual fields per year may be indicated.
Some patients are incapable of performing automated perimetry. Sometimes the patient’s vision is so poor that perimetry yields little useful information. In other patients, an adequate examination cannot be conducted because of age or physical or psychological problems. There is no point in forcing such patients to undergo perimetry repeatedly if useful clinical information is not being generated. Many patients, however, who have problems with automated perimetry at first can learn to perform adequately with proper coaching and experience.
Illuminated targets are projected onto an illuminated background. The brightness of the target (target luminance) is varied and the patient is asked to respond when the target is seen. By presenting targets that are too dim to be seen (infrathreshold) and targets that are bright enough to be seen consistently (suprathreshold), the average brightness of the dimmest test object
Figure 214-8 “Borderline” Glaucoma Hemifield Test consistent with a small difference in sensitivity between the superior and inferior hemifields in a patient with an early superior nerve fiber bundle defect. The presence of a high fixation loss rate and a low foveal sensitivity despite 20/25 vision indicates that patient reliability may be a problem. Clinically, the approach to this patient would be to evaluate the fundus carefully and repeat the field.
that can be seen is determined. This is called the threshold and is determined for multiple locations in the visual field.
The computer records the patient’s responses and the target luminance for each presentation and calculates the threshold for each test location and the statistical tests described previously.
Choice of Test Program
The standard test program used in glaucoma patients is the 30-2.  The 24-2 eliminates the peripheral test locations of the 30-2 program except for the most nasal portion of the field. Many clinicians now routinely use the 24-2 in glaucoma patients because it seems to provide as much clinically useful information as the 30-2 and saves about 5 minutes testing time per eye. The shorter testing time may reduce patient fatigue and encourage cooperation and compliance with the examination. The 10-2 program is useful in patients with very advanced field loss who have only a small island of vision persisting near fixation (see Fig. 214-9 ). The foveal sensitivity is also a very useful piece of information and should be turned “on” when performing threshold perimetry in glaucoma patients.
Humphrey has introduced a new testing algorithm called SITA (Swedish Interactive Thresholding Algorithm). SITA is available in two testing algorithms: standard and fast. The difference relates to the number of times a target is presented in order to arrive at a threshold estimation. “SITA fast” reduces testing time but is less precise and subject to greater variability. SITA
Figure 214-9 Example of the use of the central 10-2 program in a patient with end-stage glaucoma and a small residual central island of vision. The nerve fiber bundle nature of the defect is apparent at the horizontal meridian near fixation.
uses a more complex statistical technique to calculate both the threshold and the reliability indices from the patient’s responses while the test is in progress. This significantly shortens the testing time without sacrificing accuracy or affecting the ability of the test to detect abnormalities. Variability may also be reduced. SITA has largely replaced the older full-threshold programs in clinical use.   
INTERPRETATION OF RESULTS
Nerve Fiber Bundle Defects
Most visual field defects seen in glaucoma are of the nerve fiber bundle type. As a result of glaucomatous damage to ganglion cell axons at the optic nerve head, there is a loss of retinal nerve fiber bundles. This loss may be diffuse, localized or both. The characteristic shape and location of the visual field defects seen in glaucoma result from the anatomy of the retinal nerve fiber layer. The defects seen in an individual patient depend on the location of the damaged nerve fibers and whether the damage is predominantly localized or diffuse.
The most characteristic feature of the nerve fiber bundle visual field defect is the tendency to respect the horizontal meridian, especially in the nasal portion of the field. Isolated nerve fiber bundle defects rarely cross the nasal horizontal midline Typically, there is an abrupt change in sensitivity across the horizontal midline. Even in patients with more advanced visual field loss due to glaucoma there is often a detectable difference in the measured threshold on either side of the nasal horizontal midline.
Another feature of nerve fiber bundle visual field defects is the tendency to be found in the Bjerrum area, which is between
Figure 214-10 A superior visual field defect demonstrating the typical features of the nerve fiber bundle defect.
10° and 20° from fixation temporally but fans out to between 2 and 25° nasally. Scotomas in this area often assume an arcuate shape with the circumferential diameter greater than the radial diameter. Fixation itself is usually spared unless the defect is far advanced.
Clinically, nerve fiber bundle defects may appear as paracentral or arcuate scotomas, nasal steps, temporal sector defects, or various combinations ( Figs. 214-9 to 214-12 ). Any visual field defect that has nerve fiber bundle characteristics in a patient with optic disc cupping may safely be assumed to be glaucomatous in nature.
Generalized loss of retinal sensitivity, enlargement of the blind spot, and selective loss of sensitivity in the nasal periphery without specific nerve fiber bundle characteristics have been described in glaucoma. There are many other causes of these types of visual field defects. Although any of them may occur as an isolated finding in glaucoma, more commonly they are associated with a nerve fiber bundle defect.  
Artifacts Resembling Visual Field Defects
There are a number of artifacts of visual field testing that can produce results resembling those in true visual field defects. An artifact does not reflect abnormal visual function. Rather, it results from the way the patient responds to the testing situation. Generalized depressions such as are seen in patients with cataracts or small pupils are not artifacts. They are true visual field defects that reflect diminished visual function. Artifacts and nonglaucomatous visual field defects must be distinguished from each other as well as from defects due to glaucoma.
Figure 214-11 A patient with superior and inferior nerve fiber bundle defects demonstrating a paracentral scotoma superior and temporal to fixation.
The learning effect is a common artifact seen in patients undergoing their first visual field examination.   Typically, it appears as a loss of sensitivity that is most pronounced in the more peripheral portions of the field. The defect either disappears or markedly improves after the second or third examination.
An apparent depression in the superior peripheral portion of the field may resemble an arcuate scotoma in the gray scale. The superior portion of the visual field normally has lower sensitivity and higher variability.  Careful inspection of the total and pattern deviation as well as the Glaucoma Hemifield Test helps to identify this artifact. Blepharoptosis, however, even when very mild, may produce significant depressions in the superior visual field resembling the defects seen in glaucoma. Some of these defects may be quite close to fixation.
In order to obtain accurate central visual fields, the patient’s refractive correction must be placed in the perimeter. If this is not done, the visual field may appear abnormal.  In general, about 1 decibel of loss will appear for each diopter of over- or undercorrection placed in the perimeter. The loss due to incorrect refractive correction tends to be most pronounced in the central visual field.
If the pupil is smaller than 2.5?mm, an otherwise normal visual field may appear to be depressed whereas an abnormal visual field may appear worse than it really is.   Strictly speaking, this is not an artifact because the changes in the visual field result from an alteration of the visual pathway. A small pupil, however, makes interpretation of the visual field difficult. The pupil size should be recorded each time the visual field is tested. If an effect of miosis on the visual field is suspected, the pupil should be dilated prior to the examination. Many clinicians routinely dilate pupils less than 3?mm for perimetry.
Figure 214-12 A superior nasal step type of nerve fiber bundle defect. Note that in this very early defect the global indices are in the normal range, but both the Glaucoma Hemifield Test and the total and pattern deviation are sensitive enough to detect this small defect.
The rim of the lens holder may produce a scotoma by obscuring a portion of the patient’s view of the perimeter bowl ( Fig. 214-13 ). The lens holder should be placed as close to the patient’s eye as possible and the patient’s eye should be well centered behind the lens. Fatigue and an unduly long examination time may also be associated with depressed sensitivity and apparent visual field defects.
Detection of Progressive Visual Field Loss
The detection of progressive glaucomatous visual field loss with automated perimetry is an extremely complex problem that has not been satisfactorily solved.   The visual fields of both normal individuals and glaucoma patients are subject to a large degree of long-term fluctuation, which is defined as variation in measured visual thresholds of examinations performed on different days. Long-term fluctuation has been extensively studied and shown to be larger in glaucoma patients than in normal subjects, larger in more extensively damaged areas of the visual field, larger in the periphery of the field than near fixation, and larger in the superior half of the field.    
Because of the large amount of long-term fluctuation found in the visual fields of glaucoma patients, it is often difficult to decide whether the difference between two fields is due to true progressive field loss or random variation. Investigators have been grappling with this problem since the inception of automated perimetry. Even experienced clinicians often have difficulty in determining whether or not visual field defects are progressing and frequently do not agree with statistical tests applied to the visual field data. Techniques have improved agreement among clinicians and statistical tests in detecting progressive
Figure 214-13 Typical lens rim artifact simulating a dense inferior visual field defect. This completely disappeared when the patient was retested with the lens holder properly positioned.
field loss, but we still lack a generally accepted “gold standard” definition of progression that has been validated.  
Some clinical treatment trials in glaucoma have developed sophisticated techniques to detect progression. These different techniques, however, show only fair agreement when applied to the same groups of patients. Often, the different scoring and statistical techniques come to opposite conclusions about the same set of visual fields. 
Despite these problems, it has become apparent that multiple visual field examinations are required to separate fluctuation from progression and that progression is more easily detected when a series of visual fields are graphically displayed in a single printout. Clinicians should probably not even attempt to diagnose progression with fewer than four examinations and only if any changes can be documented on repeated testing.  Nor should one rely on simple inspection of individual visual field printouts.
The Humphrey software provides three ways to display serial visual field data to assist in determining the presence of progression: the Overview printout, the Change Analysis printout, and the Glaucoma Change Probability Analysis.   Another commercially available software program, PROGRESSOR, is also useful for this purpose.
The Overview printout ( Fig. 214-14 ) simply displays the gray scale, measured thresholds, and total and pattern deviations on a single sheet of paper. Trends over time may be more easily demonstrated with this printout, but no statistical analysis is performed.
The Change Analysis printout ( Fig. 214-15 ) displays serial visual field data as a set of frequency distributions of the actual measured threshold data in the form of box plots. The normal
Figure 214-14 Overview printout showing progressive glaucomatous visual field loss. The enlargement and deepening of the defect are well seen in the gray scale, total, and pattern deviation displays.
distribution is shown as a sample box plot to the left of the display. The actual values are presented in chronological order in the graph. The numbers along the Y axis of the graph represent the difference between the actual measured threshold and the age-corrected normal, in other words, the departure from normal. Positive values represent test locations with thresholds above normal and negative values, test locations with thresholds below normal. The highest value in the box plot, the top of the T, is the best point in the visual field. It is the test location with the highest value compared with the normal. The top of the box
Figure 214-15 Change Analysis printout of the patient in Fig. 214-14 showing a progressing visual field. The box plot frequency distributions demonstrate a downward trend between fields 1988 and 1997. There is a significant downward trend in both the MD and PSD plots. The last four or five fields seem to have stabilized.
is the 85th percentile. Fifteen percent of the test locations have values higher and 85% have values lower than that represented by the top of the box. The thick bar in the middle of the box is the median value; 50% of the thresholds are above and 50% below this value. The bottom of the box is the 15th percentile, and the bottom of the T is the worst point in the field. The box plot does not take into consideration the location of the test points, only their threshold values compared with normal. When a large number of fields are available, downward trends in any part of the box plot can often be easily distinguished from random fluctuation to allow a diagnosis of progression. Progression may not necessarily be due to glaucoma because cataracts, other media opacities, and retinal disease also cause progressive deterioration of the visual field.
Below the box plots, the global indices are displayed over time. Visual inspection of the MD or PSD plot over time may show progression as a consistent downward trend.
The Glaucoma Change Probability Analysis printout ( Fig. 214-16 ) displays the gray scale, total deviation, and change in threshold from baseline for each test location of a series of fields on a single sheet of paper. To the right of these data is a graph representing the probability that changes in each individual test location are outside the expected range of random fluctuation. Clear triangles represent test locations showing improvement, and solid black triangles represent test locations showing deterioration.
Progression is usually represented by a cluster of black triangles in the same location that enlarges with time. It should again be emphasized that none of these displays allows the clinician to detect progression reliably with fewer than four to six visual fields. The Glaucoma Change Probability Analysis is not available for visual fields done with SITA.
Figure 214-16 Glaucoma Change Probability Analysis printout of the patient in Fig. 214-14 . The last four field tests were done with SITA and are not included in this analysis. The first two fields serve as a baseline. The clusters of black triangles in the most recent fields identify the areas of significant progression.
Figure 214-17 A normal blue-yellow (SWAP) visual field. (Courtesy of Dr. Chris Johnson, Devers Eye Institute, Portland, Oregon.)
This chapter has emphasized the Humphrey system for automated perimetry because it is the most widely used system in North America and is the one available to the author. It should be noted that equally sophisticated testing procedures and analysis programs are available with the Octopus and other perimeters. In patients who cannot perform automated perimetry or where automated perimetry is unavailable, manual quantitative perimetry with the Goldmann perimeter is a useful alternative.
Newer types of perimetry utilizing targets more complex than simple white-on-white are now undergoing development. Although none has come into widespread clinical use, there is considerable potential for their use in the future.
The use of a blue test object on a yellow background is called short-wavelength automated perimetry or SWAP for short. This test is commercially available on the Humphrey HFA II perimeter. There is some evidence that SWAP is more sensitive than white light perimetry for detecting early glaucomatous damage, but the testing time is significantly longer and variability is increased. SWAP is probably most useful in patients with evidence of early glaucomatous change to the optic disc or nerve fiber layer but normal or borderline visual fields on standard automated perimetry ( Figs. 214-17 and 214-18 ).      
Frequency-doubling technology (FDT) is another innovation that is commercially available. FDT refers to the use of rapidly alternating dark and light stripes as a perimetric target. This produces an optical illusion of nonflickering stripes that are half as wide as the actual stripes. There is some evidence that FDT perimetry may do a better job of detecting early glaucomatous visual field loss than standard white light perimetry. The test is
Figure 214-18 A SWAP visual field showing an early superior nerve fiber bundle defect in a glaucoma patient with a completely normal white-on-white standard Humphrey field. (Courtesy of Dr. Chris Johnson, Devers Eye Institute, Portland, Oregon.)
relatively quick and considerably easier and less fatiguing for the patient. It is becoming very useful as a screening technique.    
1. American Academy of Ophthalmology. Ophthalmic procedures assessment automated perimetry. Ophthalmology. 1996;103:1144–51.
2. O’Brien C, Wild JM. Automated perimetry in glaucoma—room for improvement? Br J Ophthalmol. 1995;79:200–1.
3. Miglior S, Casula M, Guareschi M. Clinical ability of the Heidelberg retinal tomograph examination to detect glaucomatous visual field defects. Ophthalmology. 2001;108:1621–7.
4. Miglior S, Brigatti L, Lonati C, et al. Correlation between the progression of optic disc and visual field changes in glaucoma. Curr Eye Res. 1966;15:145–9.
5. Quigley HA, Tielsch JM, Katz J, et al. Rate of progression in open-angle glaucoma estimated from cross-sectional prevalence of visual field damage. Am J Ophthalmol. 1996;122:355–63.
6. Kwon YH, Kim CS, Zimmerman MB, et al. Rate of visual field loss and long-term visual outcome in primary open-angle glaucoma. Am J Ophthalmol. 2001;132:47–56.
7. Portney GL, Krohn MA. Automated perimetry background, instruments and methods. Surv Ophthalmol. 1978;22:271–8.
8. Katz J, Tielsch JM, Quigley HA, Sommer A. Automated perimetry detects visual field loss before manual Goldmann perimetry. Ophthalmology 1995;102:21–6.
9. Duke-Elder S, Ashton N, Smith RJH, Lederman M. System of ophthalmology, Vol VII, The foundations of ophthalmology. St. Louis: Mosby; 1962:393–409.
10. Duke-Elder S, Jay B. System of ophthalmology, Vol XI, Diseases of the lens and vitreous; glaucoma and hypotony. St. Louis: Mosby; 1969:469–77.
11. Lynn JR, Fellman RI, Starita RJ. Principles of perimetry. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas, 2nd ed. St. Louis: Mosby; 1996:491–521.
12. Caprioli J. Discrimination between normal and glaucomatous eyes. Invest Ophthalmol Vis Sci. 1993;33:153–9.
13. Keltner JI, Johnson CA, Quigg JM, et al. Confirmation of visual field abnormalities in the Ocular Hypertension Treatment Study. Arch Ophthalmol. 2000;118: 1187–94.
14. Heijl A, Lindgren G, Olsson J. A package for the statistical analysis of visual fields. In: Heijl A, ed. Seventh International Visual Field Symposium, Amsterdam, September 1986. Dordrecht, The Netherlands: Dr W Junk; 1987:153–68.
15. Heijl A, Lindgren G, Olsson J. Normal variability of static visual threshold values across the central visual field. Arch Ophthalmol. 1987;105:1544–9.
16. Katz J, Sommer A. A longitudinal study of the age-adjusted variability of automated visual fields. Arch Ophthalmol. 1988;105:1083–6.
17. Bebie H, Flammer J, Bebie T. The cumulative defect curve of local and diffuse components of visual field damage. Graefes Arch Clin Exp Ophthalmol. 1989;227:9–12.
18. Flammer J, Jenni F, Bebie H, et al. The Octopus G1 program. Glaucoma. 1987;9:67–72.
19. Heijl A, Lindgren G, Lindgren A, et al. Extended empirical statistical package for evaluation of single and multiple fields in glaucoma Statpac 2. In: Mills RP, Heijl A, eds. Perimetry update 1990/91. Amstelveen, The Netherlands: Kugler & Ghedini; 1991:303–15.
20. Anderson DR, Patella VM. Automated static perimetry, 2nd ed. St. Louis: Mosby; 1999:10–35.
21. Caprioli J. Automated perimetry in glaucoma. Am J Ophthalmol. 1991;111: 235–6.
22. Lynn JR, Fellman RL, Starita RI. Principles of perimetry. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas, 2nd ed. St. Louis: Mosby–Year Book; 1996: 491–521.
23. Werner EB. The normal visual field. In: Werner EB, ed. Manual of visual fields. New York: Churchill Livingstone; 1991:91–110.
24. Vingrys AJ, Demirel S. False-response monitoring during automated perimetry. Optom Vis Sci. 1998;75:513–17.
25. Heijl A, Lindgren G, Olsson J. Reliability parameters in computerized perimetry. In: Greve EL, Heijl A, eds. Seventh International Visual Field Symposium, Amsterdam, September 1986. Dordrecht: Dr W Junk; 1987:593–600
26. Bengtsson B, Olsson J, Heijl A, Rootzen H. A new generation of algorithms for computerized threshold perimetry, SITA. Acta Ophthalmol Scand. 1997;75: 368–75.
27. Katz J, Sommer A. Screening for glaucomatous visual field loss. The effect of patient reliability. Ophthalmology. 1990;97:1032–7.
28. Johnson CA, Nelson-Quigg JM. A prospective three-year study of response properties of normal subjects and patients during automated perimetry. Ophthalmology. 1993;100:269–74.
29. Bengtsson B, Heijl A. False-negative responses in glaucoma perimetry indicators of patient performance or test reliability? Invest Ophthalmol Vis Sci. 2000;41:2201–4.
30. Katz J, Sommer A, Gaasterland DE, et al. Comparison of analytic algorithms for detecting glaucomatous visual field loss. Arch Ophthalmol. 1991;109:1684–9.
31. Katz J, Sommer A, Witt K. Reliability of visual field results over repeated testing. Ophthalmology. 1991;98:70–5.
32. Anderson DR, Patella VM. Automated static perimetry, 2nd ed. St. Louis: Mosby; 1999:121–6.
33. Advanced Glaucoma Intervention Study. 2. Visual field test scoring and reliability. Ophthalmology. 1994;101:1445–55.
34. Lee M, Zulauf M, Caprioli J. The influence of patient reliability on visual field outcome. Am J Ophthalmol. 1994;117:756–61.
35. Demirel S, Vingrys AJ. Eye movements during perimetry and the effect that fixational instability has on perimetric outcomes. J Glaucoma. 1994;3:28–35.
36. Henson DB, Evans J, Chauhan BC, et al. Influence of fixation accuracy on threshold variability in patients with open-angle glaucoma. Invest Ophthalmol Vis Sci. 1996;37:444–50.
37. Flanagan JG, Wild JM, Trope GE. The visual field indices in primary open-angle glaucoma. Invest Ophthalmol Vis Sci. 1993;34:2266–74.
38. Funkhouser AT, Fankhauser F. The effects of weighting the Mean Defect visual field index according to threshold variability in the central and mid peripheral visual field. Graefes Arch Clin Exp Ophthalmol. 1991;229:228–31.
39. Anderson DR, Patella VM. Automated static perimetry, 2nd ed. St. Louis: Mosby; 1999:111–15.
40. Heijl A, Asman P. A clinical study of perimetric probability maps. Arch Ophthalmol. 1989;107:199–203.
41. Heijl A, Lindgren G, Olsson J, et al. Visual field interpretation with empiric probability maps. Arch Ophthalmol. 1989;107:204–8.
42. Asman P, Heijl A. Glaucoma hemifield test automated visual field evaluation. Arch Ophthalmol. 1992;110:812–19.
43. Asman P, Heijl A. Evaluation of methods for automated hemifield analysis in perimetry. Arch Ophthalmol. 1992;110:820–6.
44. Katz J, Quigley HA, Sommer A. Detection of incident field loss using the glaucoma hemifield test. Ophthalmology. 1996;103:657–63.
45. American Academy of Ophthalmology. Preferred practice pattern primary open-angle glaucoma. San Francisco; 1996.
46. Anderson DR, Patella VM. Automated static perimetry, 2nd ed. St. Louis: Mosby; 1999:191–245.
47. Bengtsson B, Heijl A, Olsson J. Evaluation of a new threshold visual field strategy, SITA, in normal subjects. Acta Ophthalmol Scand. 1998;76:165–9.
48. Bengtsson B, Heijl A. evaluation of a new perimetric threshold strategy, SITA, in patients with manifest and suspect glaucoma. Acta Ophthalmol Scand. 1998;76: 268–72.
49. Sekhar GC, Naduvilath TJ, Lakkai M, et al. Sensitivity of Swedish Interactive Threshold Algorithm compared with standard full threshold algorithm in Humphrey visual field testing. Ophthalmology. 2000;107:1303–8.
50. Wild JM, Pacey IE, O’Neill EC, Cunliffe LA. The SITA perimetric threshold algorithms in glaucoma. Invest Ophthalmol Vis Sci. 1999;40:1998–2009.
51. Budenz DL. Atlas of visual fields. Philadelphia: Lippincott-Raven; 1997:143–94.
52. Varma R, Minckler DS. Anatomy and pathophysiology of the retina and optic nerve. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas, 2nd ed. St. Louis: Mosby; 1996:139–76.
53. Asman P, Heijl A. Diffuse visual field loss and glaucoma. Acta Ophthalmol (Copenh). 1994;72:303–8.
54. Mutlukan E. Diffuse and localized visual field defects to automated perimetry in primary open-angle glaucoma. Eye. 1995;9:745–50.
55. Chauhan BC, LeBlanc RP, Shaw AM, et al. Repeatable diffuse visual field loss in open-angle glaucoma. Ophthalmology. 1997;104:532–8.
56. Budenz DL, Feuer WJ, Anderson DR. The effect of simulated cataract on the glaucomatous visual field. Ophthalmology. 1993;100:511–17.
57. Lindenmuth KA, Skuta GL, Rabbani R, et al. Effects of pupillary constriction on automated perimetry in normal eyes. Ophthalmology. 1989;96:1298–1301.
58. Heijl A, Lindgren G, Olsson J. The effect of perimetric experience in normal subjects. Arch Ophthalmol. 1989;107:81–6.
59. Werner EB, Adelson A, Krupin T. Effect of patient experience on the results of automated perimetry in clinically stable glaucoma patients. Ophthalmology. 1988; 95:764–7.
60. Werner EB, Krupin T, Adelson A, et al. Effect of patient experience on the results of automated perimetry in glaucoma suspect patients. Ophthalmology. 1990;97:44–8.
61. Katz J, Sommer A. Asymmetry and variation in the normal hill of vision. Arch Ophthalmol. 1986;104:65–8.
62. Meyer DR, Stern JH, Jarvis JM, et al. Evaluating the visual field effects of blepharoptosis using automated static perimetry. Ophthalmology. 1993;100:651–9.
63. Herse P. Factors influencing normal perimetric thresholds obtained using the Humphrey Field Analyzer. Invest Ophthalmol Vis Sci. 1992;33:611–17.
64. Heuer DK, Anderson DR, Feuer WJ, et al. The influence of refraction accuracy on automated perimetric threshold measurements. Ophthalmology. 1987;94:1550–3.
65. Weinreb RN, Perlman JP. The effect of refractive correction on automated perimetric thresholds. Am J Ophthalmol. 1986;101:706–9.
66. Heuer DK, Anderson DR, Feuer WJ, et al. The influence of decreased retinal illumination on automatic perimetric threshold measurements. Am J Ophthalmol.1989;108:643–50.
67. Zalta AH. Lens rim artifact in automated threshold perimetry. Ophthalmology. 1989;98:1302–11.
68. Searle AET, Wild JM, Shaw DE, et al. Time-related variation in normal automated perimetry. Ophthalmology. 1991;98:701–7.
69. Schulzer M, Anderson DR, Drance SM. Errors in the diagnosis of visual field progression in normal tension glaucoma. Ophthalmology. 1994;101:1589–95.
70. Katz J, Gilbert D, Quigley HA, Sommer A. Estimating progression of visual field loss in glaucoma. Ophthalmology. 1997;104:1017–25.
71. Katz J, Congdon N, Friedman DS. Methodological variations in estimating apparent progressive visual field loss in clinical trials of glaucoma treatment. Arch Ophthalmol. 1999;117:1137–42.
72. Boeglin RI, Caprioli I, Zulauf M. Long-term fluctuation of the visual field in glaucoma. Am J Ophthalmol. 1992;113:396–400.
73. Stewart WC, Hunt HH. Threshold variation in automated perimetry. Surv Ophthalmol. 1993;37:353–61.
74. Werner EB, Ganiban G, Balazsi AG. Effect of test point location on the magnitude of threshold fluctuation in glaucoma patients undergoing automated perimetry. In: Mills RP, Heijl A, eds. Perimetry update 1990/91. Amstelveen, The Netherlands: Kugler & Ghedini; 1991:175–81.
75. Werner EB, Petrig B, Krupln T, et al. Variability of automated visual fields in clinically stable glaucoma patients. Invest Ophthalmol Vis Sci. 1989;30:1083–9.
76. Hutchings N, Wild JM, Hussey MK, et al. The long-term fluctuation of the visual field in stable glaucoma. Invest Ophthalmol Vis Sci. 2000;41:3429–36.
77. Gloor BP, Vökt BA. Long-term fluctuation versus actual field loss in glaucoma patients. Dev Ophthalmol. 1985;12:48–69.
78. Werner EB, Bishop KI, Koelle J, et al. A comparison of experienced clinical observers and statistical tests in detection of progressive visual field loss in glaucoma using automated perimetry. Arch Ophthalmol. 1988;106:619–23.
79. Chauhan BC, Drance SM, LeBlanc RP, et al. Technique for determining visual field progression by using animation graphics. Am J Ophthalmol. 1994;118:485–91.
80. Smith SD, Katz J, Quigley J. Analysis of progressive change in automated visual fields in glaucoma. Invest Ophthalmol Vis Sci. 1996;37:1419–28.
81. Fitzke FW, Hitchings RA, Poinoosawmy D, et al. Analysis of visual field progression in glaucoma. Br J Ophthalmol. 1996;80:40–8.
82. Katz J. Scoring systems for measuring progression of visual field loss in clinical trials of glaucoma treatment. Ophthalmology. 1999;106:391–5.
83. Morgan RK, Feuer WJ, Anderson DR. Statpac 2 glaucoma change probability. Arch Ophthalmol. 1991;109:1690–2.
84. Viswanathan AC, Fitzke FW, Hitchings RA. Early detection of visual field progression in glaucoma: a comparison of PROGRESSOR and STATPAC2. Br J Ophthalmol. 1997;81:1037–42.
85. Johnson CA. Early losses of visual function in glaucoma. Optom Vis Sci. 1995;72:359–70.
86. Johnson CA, Brandt JD, Khong AM, Adams AJ. Short-wavelength automated perimetry in low, medium and high risk ocular hypertensive eyes. Arch Ophthalmol. 1995;113:70–6.
87. Felius J, de Jong LAMS, van den Berg TJTP, et al. Functional characteristics of blue-on-yellow perimetric thresholds in glaucoma. Invest Ophthalmol Vis Sci. 1995;36:1665–74.
88. Kwon YH, Park HJ, Jap A, et al. Test-retest variability of blue-on-yellow perimetry is greater than white-on-white perimetry in normal subjects. Am J Ophthalmol. 1998;126:29–36.
89. Mansberger SL, Sample PA, Zangwill L, Weinreb RN. Achromatic and short-wavelength automated perimetry in patients with glaucomatous large cups. Arch Ophthalmol. 1999;117:1473–7.
90. Girkin CA, Emdadi A, Sample PA, et al. Short-wavelength automated perimetry in the detection of progressive optic disc cupping. Arch Ophthalmol. 2000;118: 1231–6.
91. Blumenthal EZ, Sample PA, Zangwill L, et al. Comparison of long-term variability for standard and short-wavelength automated perimetry in stable glaucoma patients. Am J Ophthalmol. 2000;129:309–13.
92. Caprioli J. Should we use short-wavelength automated perimetry to test glaucoma patients? Am J Ophthalmol. 2001;131:792–4.
93. Johnson CA, Samuels SJ. Screening for glaucomatous visual field loss with frequency doubling perimetry. Invest Ophthalmol Vis Sci. 1997;38:413–25.
94. Trible JR, Schultz RO, Robinson JC, Rothe TL. Accuracy of glaucoma detection with frequency-doubling perimetry. Am J Ophthalmol. 2000;129:740–5.
95. Burnstein Y, Ellish NJ, Magbalon M, Higginbotham EJ. Comparison of frequency doubling perimetry with Humphrey Visual Field analysis in a glaucoma practice. Am J Ophthalmol. 2000;129(3):328–33.
96. Cello KE, Nelson-Quigg JM, Johnson CA. Frequency doubling technology perimetry for detection of glaucomatous visual field loss. Am J Ophthalmol. 2000;129:314–22.
97. Landers J, Goldberg I, Graham S. A comparison of short-wavelength automated perimetry with frequency doubling perimetry for the early detection of visual field loss in ocular hypertension. Clin Exp Ophthalmol. 2000;28:248–52.