Perimetry is a vital part of the armamentarium available to the ophthalmologist for the diagnosis and monitoring of glaucoma. It is important to be thorough with the basic principles of perimetry, not only to interpret any visual field but also to order the correct test for an individual patient.
The user friendly interface provided by automated perimeters have made them immensely popular for diagnosis and monitoring of the visual field. Advantages of automated perimetry include reproducibility, progression monitoring, standardized test formats, facility for data storage. In this chapter, we will cover the essential aspects of Humphrey Field Analyser (HFA), which is the most widely used machine worldwide.
Normal Visual Field
Traquair described the normal visual field as an island of vision in sea of darkness.1 In the normal visual field examination, the fovea is the most sensitive point tested and represents the peak. The island of vision extends roughly 60 degree superiorly and nasally, 75 degree inferiorly, and 100 degree temporally2 (Fig. 1.1).
The hallmark of glaucoma is the nerve fiber bundle defect that results from damage at the optic nerve head. The visual field defect is seen in the areas subserved by the damaged nerve fibre bundle corresponding to the ONH damage. The superior and inferior poles of the optic nerve are more susceptible to glaucomatous damage. However damage to small scattered bundles of optic nerve axons also occur, resulting in generalized decrease in sensitivity.
The typical pattern of glaucomatous visual field loss at the superior and inferior poles may be attributed to following. The neuroretinal rim is physiologically broader at the inferior and superior poles than at the nasal and temporal poles.3 The lamina cribrosa shows larger pores and a higher ratio of pore to interpore connective tissue area in the inferior and superior regions as compared with the temporal and nasal regions. A high ratio of pore area to total area is considered to predispose to glaucomatous nerve fiber loss.4-7 Glaucomatous backward bowing of the lamina cribrosa to the outside, mainly in the inferior and superior regions, has been shown on scanning electron microscopic photographs of glaucomatous eyes.8
The lamina cribrosa is thicker in the disc periphery, where the nerve fiber bundles have a slightly more bent course through the lamina cribrosa9 and where they are lost earlier than in the center of the optic disc (Fig. 1.2).
Principles of Visual Field Testing
The human visual system is more adapted to perceive contrast rather than absolute magnitude of light. Estimating differential light sensitivity of stimulus against a constant illuminated background is the central essence of static perimetry.10 VF testing explores differential light sensitivity throughout field of vision. Visibility of a stimulus depends on its intensity, duration, size and color, background intensity, attentiveness of patient and refractive status of eye.11
Threshold is defined as the stimulus intensity at which 50% of the presented stimuli are perceived by the patient. Suprathreshold stimuli are seen more than 50% of the time whereas infrathreshold or subthreshold stimuli are seen less than 50% of the time. Plotting the probability of seeing a stimulus against the range of stimulus intensity gives the frequency of seeing curve (Figs 1.3 and 1.4).
Light Intensity and Its Units
Asb = unit of luminance (Absolute unit)
1 asb = 0.1 miliambert = 1 lumen of total emittance/sq.m
1 candela = 1 lumen/steradian
1 asb = 1 pi cd/sq.m.
The attenuation of light is expressed in 1/10th of log units and represents change relative to maximal stimulus intensity. Hence decibel (dB) is a relative unit and different on different instruments. The dB value refers to the retinal sensitivity rather than the stimulus intensity, with 0 dB corresponding to maximum brightness. A 3dB change is equivalent to doubling of intensity, and when judging visual field loss, represents a significant change in visual field sensitivity. The maximum brightness that the Humphrey perimeter can produce is 10000 asb. Weakest stimulus which can be perceived in healthy young patients is 40 dB, i.e. 1/10000th of maximal stimulus. Perception of even weak stimuli, i.e. in range of 40–50 dB should alert the glaucomatologist of false positive responses.
The geographic pattern of visual field loss is more important to diagnose glaucoma rather than threshold determination.
The salient points which require attention for choosing correct test include the test strategy (supra/threshold), test pattern (30-2, 24-2, 10-2, macular program,-1 strategies), threshold determination (bracketing).
Estimating threshold using full threshold strategy requires the threshold to be crossed twice. If the initial stimulus is not perceived by the patient (subthreshold stimulus), intensity of stimulus is increased by 4dB until the patient perceives it (first crossing of threshold). The stimulus intensity is then decreased in steps of 2 dB until the patient stops perceiving the stimuli. At this point the visual threshold is crossed again. The Humphrey Visual Field Analyzer reports threshold values as the last seen stimulus using the 4–2 strategy. If the initial stimulus is suprathreshold, stimulus intensity is decreased by 4-dB steps until the threshold is crossed, then increased in 2-dB steps (the threshold again is doubly crossed). Threshold estimation requires approximately five stimuli per test location in full threshold strategy. Patients should be told that they will miss 50% of stimuli and many of the stimuli will be too dim (Fig. 1.5).
Initial Point of Threshold Estimation
The strategies used to estimate threshold are continuously evolving in the search of a method which would give accurate threshold estimation in shortest time. Time interval for threshold estimation may be decreased if starting intensity is close to the actual threshold. Starting intensity can be adjusted according to age expected normal, from threshold. The initial presentation may be determined in part from an expected normal visual sensitivity, from threshold estimate at nearby test point, or from the threshold recorded on a previous test on that patient. But this estimation may bias the final outcome.
The approximate time taken for each program of visual field testing is represented in Table 1.1.
Swedish Interactive Threshold Algorithm (SITA)
The SITA has largely replaced full threshold testing. Two SITA algorithms are currently available, SITA Standard and SITA Fast, which are analogous to the full threshold and FASTPAC algorithms, respectively. SITA uses visual field modeling based on frequency-of-seeing curves of glaucoma and normal patients. The test procedure starts by measuring threshold values at four primary points, one in each quadrant of the field at 9 degree from the fixation. Starting from initial stimulus intensity of 25 dB, stimuli are altered in 4–2 dB step size. Threshold values obtained at these primary points are then used to calculate starting levels at adjacent points. As the test continues, last seen stimulus intensities in neighboring points are used to calculate starting values in new points not previously opened for testing. Stimuli are presented in pseudorandom order. Stimulus used in SITA is always size III. Patient response time is taken into account to decide the interval between the successive stimuli.
An alternative threshold algorithm for the HFA, the FASTPAC strategy, uses a single crossing of threshold with a constant 3 dB step size, and threshold is designated as the last-seen stimulus luminance.
The examination duration of the FASTPAC algorithm is approximately 35% of that of the full threshold algorithm but is at the expense of an approximate 25% increase in the intratest variability (short-term fluctuation) and an underestimation of focal loss in glaucoma.
By definition, short term fluctuation at a point is standard deviation around the mean of replicate measurements. Global index of short-term fluctuation is estimated by retesting ten preselected sample of locations (Fig. 1.6).
Threshold fluctuation may be seen with eccentricity from fixation, reduced retinal sensitivity, due to learning effects, reliability, pupil size, age and mode of stimulus presentation.11 SF as a global index was abandoned in some recent strategies when it became obvious that the SF estimate depended on whether the sampled areas were normal or abnormal and therefore SF is different at various locations.12
Long-term fluctuation is the actual variation of threshold sensitivity due to physiological factors, irrespective of state of alertness. It should be taken into account when interpreting visual field progression and field must be retested twice before confirming progression.
Choosing Test Pattern
Commonly used test patterns in Humphrey field analysis include 24-2, 30-2, 10-2, and macular function test (Table 1.1).
The 30-2 algorithm samples 76 test points with a uniform 6 degree paraxial grid within 30 degree from fixation which is offset from horizontal and vertical meridian. All Humphrey programs ending with-2 are paraxial. The 30 degree limitation is appropriate because this area is affected in almost all cases of visual field disturbance. The central 30 degree area also represents about 60% of all nerve fibers. The pattern of 30-1 is axial whereas pattern of 30-2 and 24-2 is paraxial. The-1 algorithmis no longer used clinically.
24-2 algorithm tests 54 locations (excludes outer outlined locations of 30-2 as shown in Figure 1.7) extending to 21 degree superiorly, inferiorly, and temporally but tests to 27 degree nasally to involve the nasal locations of 30-2. It thus decreases the test time considerably without compromising much on the data accuracy.
10-2 pattern tests 68 points in central 10 degree with a 2 degree grid offset 1 degree from the vertical and horizontal meridian. They are useful in defining central and paracentral scotomas, and are commonly used in advanced glaucoma (Fig. 1.8).
Macular test tests retinal sensitivity in central 5 degree. 16 points are tested, and threshold is estimated thrice at each location to estimate short term fluctuations (Fig. 1.9).
Stimulus size doubles in diameter and increases in area by a factor of four progressively with increasing Goldmann size as shown in Table 1.2. Doubling the stimulus diameter has roughly the same effect as increasing intensity by 5 dB. In automated static perimetry, size III (4 mm²)is selected to permit visibility determinations at diseased areas. Size I stimulus was originally used by Goldmann for threshold determination in healthy eyes. Size I stimuli require sharp focusing of retinal image. Larger sizes up to size V (64 mm²) can be selected in patients with central scotoma or macular disease.
The principle of temporal summation states that visibility of stimulus increases with duration of exposure upto a critical time period, usually 100 ms. Beyond this time, there is no further increase in stimulus visibility irrespective of stimulus duration. Humphrey perimeter uses a stimulus duration of 200 ms which is long enough for visibility to be affected by small variations in duration, but still shorter than the latency for voluntary eye movements (about 250 ms).10
As differential light is measured, i.e., white stimulus of given intensity is less visible against background of similar intensity than against a dark one. Under photopic conditions, visibility depends on contrast (Weber's law).
Background intensity also determines light adaptation of retina which influences visibility. The illumination of background is kept in low photopic range, i.e. 31.5 asb commonly, because dimmer illumination tends to flatten and depress the retinal sensitivity curve centrally.13 Profound dark adaptation may cause relative physiologic central scotoma due to poor response from cones.
Measures gaze direction with a precision of about one degree and records a measurement each time a stimulus is presented. The gaze tracking results are shown on the video screen during testing and are printed at the bottom of the results printout. On the gaze printout, lines extending upward indicate the amount of gaze error at each stimulus presentation, with full scale indicating gaze errors of 10 degrees or more. Lines extending downward indicate that the instrument was unsuccessful in measuring gaze direction during that particular stimulus presentation, e.g. blink.
Confounding Factors and Artefacts
Pupil size: Pupil size less than 3 mm if combined with lens opacity produces diffuse depression of visual fields (Fig. 1.10). Hence pupil size should be recorded on each test, when considering any change in field test result. Patient having miosis due to neuro-ophthalmic diseases, senile miosis, or miosis due to use of regular miotics should be dilated prior to testing.
Contrast sensitivity and glare: Any opacity in the media may cause scattering of light and decreased contrast which may result in reduction of stimulus visibility.
Refractive status of the eye: One diopter of refractive blur in an undilated patient will produce a little more than one decibel of depression of hill of vision when testing with a Goldmann size 3 stimulus.14 The nominal testing distance of the Humphrey HFA II perimeter is 30 cm and fully presbyopic patient is provided with 3.25 diopter near additions relative to their distance refraction. Patients who are less than fully presbyopic are given smaller additions. Usually all refractive corrections are accomplished using standard 37 mm trial lenses held in place by a trial lens holder attached the perimeter, but correction may be done with the patients’ own spectacles as long as they are single lenses or contact lenses. Larger stimuli are less affected than smaller stimuli by refractive errors.
Learning effects: Learning effects are more in periphery, more in fields with moderate loss, and more in points with borderline sensitivity (Fig. 1.11). Learning effects are also task specific, i.e. familiarity to SAP doesn't eliminate learning effect in SWAP.
Fatigue: Longer test duration and ill health can lead to fatigue during test which may manifest as high false negatives and clover leaf pattern on gray scale (Fig. 1.12).
Lens artifact: Lens with rim kept away from the eye may result in ring scotoma (Fig. 1.13). Decentred lens can cause arcuate scotoma. This can be avoided by using contact lens or rimless lenses, and correct centration and positioning of lens.
Lid artifact: Drooping of upper lids may cause superior arcuate defects or superior nasal step, and can be avoided by taping the upper lid before performing the test.
False positives (FPs): Occasionally there are intervals during which the machine makes a soft click but shows no target. This error happens when an anxious subject presses the button during this interval. Gray scale of subject with high false positive appears abnormally white seen as “white scotomas”. In SITA strategy the number of these anticipatory responses is counted which are made too soon after a flash to be a response to the light.
False negatives (FNs): A fairly bright suprathreshold target is flashed in a region previously tested with fainter targets. If the patient fails to indicate its presence, this is an FN error. A high FN rate usually implies inattention or fatigue and will be accompanied by a field with scattered factitious elevations of threshold.
False positive and false negative error ≥33% are a warning of poor reliability and are indicated by an “XX” beside the aberrant value and a printed statement of “low patient reliability,” in the upper left corner, above the GHT test.
Figure 1.14 shows a field with high false positive and false negative error showing an abnormal field which became normal in the subsequent field once reliability indices improved.
Vision is a combination of several, distinct quantifiable functions including visual acuity, color vision, vernier (alignment) acuity, the perception of movement and change in luminous intensity (flicker) or differences in luminous intensity (contrast). Visual fields are a composite of all these attributes. From the time that hemianopsias were first recognized by Hippocrates, more than two thousand years ago, perimetry has evolved from confrontation visual field evaluation to the current state of the art perimeters which even offer structural and functional correlates. The wealth of information thus made available from visual field testing has revolutionized current glaucoma practice, and has provided new dimensions to the management of neurological disease as well.
The authors gratefully acknowledge Dr Anuj Ponnappa and Dr Muazzam Ali Akbar for their contribution to this chapter.
- Traquair HM. Clinical detection of early changes in the visual field. Trans Am Ophthalmol Soc. 1939;37:158-79.
- From Allingham RRE, et al. Shields’ textbook of glaucoma. 5th edn. Lippincott Williams and Wilkins; 2004.
- Jonas JB, Gusek GC, Naumann GOH. Optic disk, cup and neuroretinal rim size, configuration, and correlations in normal eyes. Invest Ophthalmol Vis Sci. 1988;29:1151-8. 1991:32:1893.
- Jonas JB, Mardin CY, Schlötzer-Schrehardt U, Naumann GOH. Morphometry of the human lamina cribrosa surface. Invest Ophthalmol Vis Sci. 1991;32:401-5.
- Minckler DS, McLean IW, Tso MOM. Distribution of axonal and glial elements in the rhesus optic nerve head studied by electron microscopy. Am J Ophthalmol. 1976;82:179-87.
- Minckler DS. The organization of nerve fiber bundles in the primate optic nerve head. Arch Ophthalmol. 1980;98:1630-6.
- Radius RL. Regional specificity in anatomy at the lamina cribrosa. Arch Ophthalmol. 1981;99:478-80.
- Quigley HA, Addicks EM, Green WR, Maumenee AE. Optic nerve damage in human glaucoma: II. The site of injury and susceptibility to damage. Arch Ophthalmol. 1981;99:635-49.
- Dichtl A, Jonas JB, Naumann GOH. Course of the nerve fiber bundles through the lamina cribrosa. Graefes Arch Clin Exp Ophthalmol. 1996;234:581-5.
- Anderson DR, Patella VM. Automated static perimetry. 2nd edn. Mosby; 1999.
- Stewart CW, Hunt HH. Survey of Ophthalmology. volume 37. number 5. march-april 1993.
- Anderson/Ophthalmol Clin N Am. 2003;16:205-12.
- Kramer SG, Survey of ophthalmology. 1991;36(1).
- Weinreb RN, Perlman JP. The effect of refractive correction on automated perimetric thresholds. Am J Ophthalmol. 1986; 101:706-9.