Decreased Visual Field Sensitivity Measured 1 Day, Then 1 Week, after Migraine

Abstract

purpose. To determine whether perimetric performance is worse the day after a migraine than prior interictal measurements, and if so, to determine whether differences have resolved by 1 week after migraine.

methods. Twenty-two nonheadache control subjects (aged 18–45 years) and 22 migraineurs (aged 18–45 years: 10 migraine with visual aura, 12 migraine without aura) participated. Standard automated perimetry (SAP) and temporal modulation perimetry (TMP) were measured by perimeter (model M-700; Medmont, Pty Ltd., Camberwell, Victoria, Australia). Control subjects attended two test visits: baseline and retest. Migraineurs attended three times: baseline (≥4 days after migraine), the day after the offset of the next migraine, and 7 days later. Groups were compared using the global indices of the perimeter: Average Defect (AD) and Pattern Defect (PD), in addition to point-wise comparisons.

results. Group migraineur TMP performance was significantly worse the day after a migraine, showing decreased general sensitivity and increased localized loss. Performance measured 7 days later was not significantly different from that measured the day after a migraine. Group migraineur SAP performance was not significantly worse after migraine; however, a subgroup of six eyes from five patients had 10 or more visual field locations with decreases in sensitivity greater than control test–retest 95% confidence limits.

conclusions. Decreased visual field performance was present after migraine, as well as greater test–retest variability in the migraine group compared with control subjects. As migraineurs constitute 10% to 15% of the general population, the presence of this subgroup of patients with periodic prolonged decreased visual field sensitivity after migraine has implications for differential clinical diagnosis, and for clinical research using perimetry.

People with migraine have increased prevalence of visual field deficits relative to nonheadache-suffering control subjects, when measured at times between migraine events. Visual field deficits in migraineurs have been demonstrated using a variety of different perimetric strategies including white-on-white standard automated perimetry (SAP), ^{1 2 3} temporal modulation perimetry (TMP), ⁴ and short-wavelength automated perimetry (SWAP), ⁵ suggesting deficits in migraineurs that are not neural pathway specific. Whereas the symptomatology of migraine is largely cortical, the visual field deficits are often unilateral, and sometimes arcuate, consistent with involvement of the precortical visual pathways in some individuals. ^{1 4 5} Of possible relevance is the observation that migraine is more common in people with glaucoma than in the general population ^{6 7 8 9} ; however, this finding has not been universal. ^{10 11} Migraine has been proposed as a vascular risk factor for glaucoma, particularly normal tension glaucoma (NTG), ¹² and has recently been identified as an important risk-factor for progression of visual field deficits in people with NTG. ¹³ Although there is a putative link between migraine and glaucoma, the relationship, if any, between interictal visual field deficits in healthy young migraineurs and the risk of development of eye disease is not known. 

Migraine is a very common condition affecting between 10% and 15% of the population. ¹⁴ Previous studies indicate that between 30% and 60% of these individuals have visual field deficits at times between migraines. ^{1 3 5 15} Given that the prevalence of glaucoma is approximately 3% of the population over the age of 50, ^{16 17} and that estimates of the percentage of the glaucoma population with a history of migraine is, at most, 30%, ⁶ it is evident that not all healthy young migraineurs with interictal visual field abnormalities will progress to having glaucoma. Indeed, we do not know whether the presence of visual field deficits between migraines is predictive of future eye disease. It is not known whether migraine events themselves affect visual processing, possibly in a cumulative fashion, or whether there are underlying vascular, metabolic, or neurologic differences in some migraineurs that predispose them to both migraine and to eye disease. 

Most previous perimetric studies of migraineurs have measured fields at a single time-point cross-sectionally. ^{1 2 4 5} Drummond and Anderson ¹⁸ demonstrated peripheral visual field constriction with kinetic targets, in a group of subjects with migraine with aura the day after a migraine, that was largely resolved at 7 to 10 days after migraine, but they did not compare performance to test–retest variability in a comparable group of nonheadache control subjects. Their study did not find deficits in individuals who had migraine without visual aura. However, the perimetric task used was not capable of identifying small, localized deficits. ¹⁸ In contrast, we have followed the perimetric performance of one subject with migraine without aura who had a localized TMP deficit that largely but not completely resolved over a 30-day period after migraine. ⁴ Sullivan-Mee and Bowman ¹⁹ also documented two case reports of migraineurs with persistent visual field deficits (which they define as present >10 days after migraine) that gradually resolved. The findings of Drummond and Anderson ¹⁸ suggest that such persistence of visual field deficits is likely to be rare. 

Studying the time course of visual field deficits relative to migraine events may provide information essential to understanding their relevance, if any, to eye disease, and for enabling correct differential diagnosis of visual field loss due to treatable cause (such as glaucoma). For example, if visual field deficits are directly related to migraine events, therapy to minimize migraine occurrence may have future benefit in reducing the risk of eye disease in susceptible individuals. Even if migrainous visual field deficits have no long-term cumulative significance, knowing whether duration after migraine is likely to affect visual field performance is essential for accurate perimetric interpretation for the management of other eye disease in migraineurs. Our present study was motivated by a need to more systematically investigate postmigraine visual field performance, with comparison to test–retest variability in nonheadache control subjects. Migraineurs were assessed interictally (4 days or more after migraine) then at 1 day and 7 days after the cessation of symptoms of their next migraine. Testing at 1 and 7 days after migraine was chosen to enable comparison with the study of Drummond and Anderson. ¹⁸ Although it would be of interest to test at longer durations after migraine, we wanted to include a representative sample of typical migraineurs, which includes people whose have migraines as frequently as once a week. Two perimetric tasks were used: SAP and TMP. SAP was included, as it is the most commonly used clinical visual field task, and several studies have demonstrated SAP visual field deficits in people with migraine. ^{1 2 3} TMP was also included, as we have previously demonstrated deficits with TMP that were not measurable using SAP. ⁴  

Methods

Subjects

Twenty-two nonheadache control subjects (aged 18–45 years) and 22 individuals with migraine (aged 18–45 years) participated. Subjects were approximately, but not exactly age matched, and there was no significant difference between the ages of the groups (t(42) = 1.22; P = 0.23). Subjects were required to meet the following visual and ocular health criteria: best corrected visual acuity of 6/7.5 or better, refractive errors less than ±5.00 D sphere and ±2.00 D astigmatism, normal anterior eye and ophthalmoscopic examination, intraocular pressure less than 21 mmHg, no history of diabetes or other systemic disease known to affect ocular function, with the exception of migraine, and no currently prescribed medications known to affect visual field sensitivity or contrast sensitivity. Migraine subjects were required to have migraine symptoms meeting the International Headache Society criteria ²⁰ for either Migraine with Aura (MA) or Migraine without Aura (MO). Ten of the subjects had MO and 12 MA. All MA subjects had visual symptoms as part of the aura. Control subjects were required to have had fewer than four headaches in the past year and to have never experienced a migraine. All subjects provided written informed consent in accordance with a protocol approved by the University of Western Australia Human Research Ethics Committee and in agreement with the tenets of the Declaration of Helsinki. 

Perimetric Tests

SAP and TMP were performed with a perimeter (model M700; Medmont Pty Ltd., Camberwell, Victoria, Australia). A detailed description of the perimeter can be found elsewhere. ²¹ The Medmont SAP and TMP tasks have recently been shown to yield results comparable to those returned by the Humphrey Field Analyser (HFA; Carl Zeiss Meditec, Dublin, CA) and Humphrey-Zeiss frequency doubling perimeter. ²² In brief, the perimeter uses 0.43° (Goldmann size III) light-emitting diodes (LEDs) as stimuli (λ_max = 565nm). The bowl luminance is 3.2 cd/m² (CIE1931 x: 0.53, y: 0.42) and the maximum stimulus luminance is 320 cd/m². SAP thresholds were assessed using the Central Threshold test, which uses a zippy estimation by sequential testing (ZEST) thresholding procedure. ²³ Flicker thresholds were determined with the Auto-Flicker test. This test presents luminance pedestal flickering stimuli, with the flickering component of the stimulus modulated around a pedestal luminance, that differs from the background luminance. The temporal frequency of the stimuli is varied with eccentricity to improve the dynamic range of the test (1°–3°, 18 Hz; 6°, 16 Hz; 10°–15°, 12 Hz; and 22°, 9 Hz). Thresholds are measured with a 6/3-dB staircase, and stimuli are presented for 800 ms. 

The test patterns for the SAP and TMP tests differ slightly. Figure 1 provides an example of the visual fields measured in one migraine subject, and demonstrates the test pattern. Test stimuli are arranged in concentric rings. Thresholds are collected from the 3°, 6°, 10°, 15°, 22° and 30° rings for SAP (see Fig. 1a ) and the 1°, 3°, 6°, 10°, 15°, and 22° rings for the flicker test (see Fig. 1c ). There are 99 test locations for SAP and 73 test locations for TMP. 

Visit Schedule

Control subjects participated in two test visits: baseline and retest. Both perimetric tests were performed at each visit. Test visits were approximately 1 hour in duration. Migraine subjects attended a baseline test visit at four or more days after migraine. Subsequent test visits were scheduled for 1 day after the cessation of migraine symptoms and then 1 week after the same migraine. 

Visual Field Analysis

Analysis was performed using the global indices returned by the perimeter and on a point-wise basis. The Medmont perimeter returns two global indices: the Average Defect (AD), and the Pattern Defect (PD) index. The AD indicates whether there is a generalized depression or elevation across the visual field, compared with the internal normative database of the perimeter, and this index is calculated in a manner similar to the Total Deviation index of the HFA. The PD index is similar to the Pattern Standard Deviation index of the HFA, where higher PD values indicate local asymmetries in an individual’s visual field relative to the remainder of their visual fields. 

For all point-wise analyses, the locations immediately above and below the blind-spot were excluded for SAP, as these may encroach on the physiological blind spot. These locations are not included in the TMP test pattern (see Fig. 1c ). Statistical comparisons were performed on computer (SigmaStat, ver. 3.0; SPSS Science, Chicago, IL). 

Results

Duration after Migraine at the Baseline Visit

The baseline visit was scheduled four or more days after the end of a migraine for the participants in the migraine group. The duration ranged from 4 to 120 days (Fig. 2) , with most subjects being tested more than 7 days after migraine. 

Visual Field Performance at the Baseline Visit

Figure 3 shows the AD and PD indices for both SAP and TMP at the baseline and retest visits (1 day after migraine retest for migraineurs). Figure 3 shows pooled data, and includes both eyes of both migraine groups. Pooling was justified by nonsignificant results (P > 0.05) returned by analysis with two-way repeated measures ANOVA in which the factors were eye, and migraine group. A separate ANOVA was run for AD and PD indices for both SAP and TMP, and no significant difference in performance between eyes or migraine groups was found for either the test or global index. 

Figures 3a and 3b show the data for SAP. At baseline, the migraine group showed significantly lower generalized sensitivity than control subjects (two-way, repeated-measures ANOVA; AD: F_42,1 = 4.10; P = 0.05) but no group differences in PD index (F_42,1 = 1.14; P = 0.29). Figures 3c and 3d show the distribution of global indices for TMP. At baseline, group comparisons of the TMP global indices approached significance (two-way, repeated-measures ANOVA; AD: F_42,1 = 3.23; P = 0.08; PD: F_42,1 = 2.49; P = 0.12). However, several migraine subjects demonstrated markedly abnormal PD indices (five eyes from four subjects were flagged at the P < 0.01 level relative to the Medmont normative database) and these subjects are represented as the outliers, and the 90% percentile whisker, in Figure 3d , left. These subjects demonstrated localized visual field deficits in response to flickering stimuli. 

Control Retest Performance

In addition to the baseline distribution of the global indices for the control population, Figure 3 also shows the retest distribution. Paired t-tests demonstrated no significant difference in control group performance for SAP at baseline versus retest visits for either the AD (t(43) = 0.52; P = 0.60) or PD (t(43) < 0.01; P > 0.99) indices. Likewise, there was no significant difference in control group global indices for TMP between baseline and retest visits (AD: t(43) = 0.52; P = 0.60; PD: t(43) = −0.07; P = 0.95). Consequently, there was no evidence of a significant learning effect for the control group on either perimetric task. 

Migraineur Performance 1 Day after a Migraine

Subjects were required to keep a headache diary and record their antimigraine medications. These included over-the-counter (OTC) analgesics (aspirin, paracetamol), OTC NSAIDs, and antiemetics. These common medications are not known to cause visual field deficits, but to minimize any possible medication effects, we tested migraine participants 1 day after the cessation of their migraine symptoms to allow for medication washout. 

The distribution of global indices for the migraine participants obtained the day after a migraine is shown in the right-hand panels of Figure 3 . For SAP (Figs. 3a 3b) visual inspection of Figure 3a shows a similar distribution of AD at baseline and retest visits (paired t-test, t(43) = 0.19; P = 0.85). The distribution of the PD index for SAP is wider at the after-migraine visit than at baseline, with the worst performing subjects being markedly abnormal (four eyes of three subjects were flagged at P < 0.01 relative to the Medmont normative database and are represented as the outliers in Fig. 3b ); however, there is not a significant group difference (paired t-test, t(43) = −1.2, P = 0.23). 

The migraine group’s TMP performance, measured 1 day after migraine, was significantly worse than at baseline for both the AD and PD indices. For AD, there was a small but statistically significant decrease in generalized sensitivity across the field (paired t-test: t(43) = 3.2, P < 0.01: mean at baseline = −1.13 dB; mean at 1 day after migraine = −1.55 dB). A significant decrease of almost 3 dB in the median PD indicates an increased presence of local field asymmetries relative to baseline in the migraine group (Wilcoxon Signed Rank Test: P < 0.01; median at baseline = 1.68 dB, median at 1 day after migraine = 4.53 dB). 

Inspection of Individual Test–Retest Performance as a Function of Deficit Severity

Visual inspection of the right-hand panels of Figure 3 demonstrates that some migraineurs returned markedly abnormal PD indices at the test visit 1 day after migraine. However, it is not clear from Figure 3 whether these are subjects that had milder deficits at baseline or were subjects with completely normal baseline performance. To explore this question, Figures 4 and 5 show scatterplots of the global indices returned at baseline and retest for SAP and TMP. Both eyes of each subject are included. Migraine subjects with baseline postmigraine durations of 1 week or less (five subjects, open circles) were individually numbered consistent with Figure 2 (Fig. 3b 3d) . The four subjects with the longest durations after migraine at baseline are also identified individually as the squares. Symbols appearing in the shaded areas are eyes where retest performance was worse than at baseline. Figures 4a and 4b show that generalized sensitivity for SAP was similar for both baseline and retest for most control and migraine subjects. In the migraine group, two eyes (appearing in the bottom right corner of Fig. 4b ) had a large reduction in AD at retest relative to baseline (3.72 and 5.20 dB). Figures 4c and 4d demonstrate a wider scatter in baseline–retest performance for the PD index for both migraine and control groups. 

Figure 5 shows scatterplots of baseline versus retest global indices (1 day after migraine retest for migraineurs) for TMP. Figure 5b demonstrates that most migraine participants showed a small generalized decrease in sensitivity relative to baseline. In addition to this small reduction in general field sensitivity, some subjects showed large increases in PD index the day after a migraine (Fig. 5d) . Some of these subjects had global indices within the normal range at baseline, whereas others demonstrated lesser deficits at baseline. 

Inspection of the individually identified subjects in Figures 4 and 5 shows that duration after migraine at baseline did not predict performance in these subgroups—performance was scattered and overlapped. There was no trend in those subjects with shortest duration after migraine at baseline to show least change after the migraine event. 

Number of Visual Field Locations with a Significant Decrease in Sensitivity at Retest for Individual Migraine Subjects

The global perimetric indices shown in the previous figures permit a summary of visual field performance for each individual, but do not provide information regarding which visual field locations or how many visual field locations demonstrated decreased sensitivity. To explore these questions, as well as whether visual field deficits were unilateral or bilateral in individual subjects, individual migraine subject performance was compared on a point-wise basis with that of the control group. The baseline–retest difference was determined for control subjects for each individual location in the field, for the two test procedures. From these, we determined two-sided 95% confidence limits of control variability. For each migraine subject, we determined the number of locations with baseline–retest scores outside the location-specific confidence interval. These are plotted in Figure 6 , where white bars indicate an improvement in performance and black bars a decrease. If a visual field had more than six locations for SAP or five locations for TMP with a significant decrease (or increase) in sensitivity, it was considered statistically significant at P < 0.05 (derived as described elsewhere, ⁵ and shown by the dotted lines in Fig. 6 ). The subject numbering is consistent with that in Figure 2 , with subjects 1 to 10 being from the migraine-without-aura group. Although were no significant group differences between subjects with and without visual aura on any of the measures in this study, Figure 6 enables visual comparison of individuals in the two subject groups. Figure 6 also enables comparison of individual subject performance across the two perimetric tasks. 

Figures 6a and 6b demonstrate that whereas most subjects returned similar SAP results at test and retest, a subgroup of migraineurs had a significant number of locations with a greater than expected decrease in sensitivity, given the test–retest variability in the control group. Six eyes of five subjects had 10 or more visual field locations identified. A minority of subjects demonstrated a significant number of locations with an improvement of sensitivity relative to baseline. The results are similar for TMP (Figs. 6c 6d) . Figure 6 does not reveal a trend for differential test–retest variability performance on either SAP or TMP in subjects with or without visual aura. Whereas some subjects had a significant number of locations with worse performance the day after a migraine compared with baseline in both eyes, visual inspection of the fields revealed arcuate deficits in most cases. None of the migraine subjects demonstrated a homonymous hemianopic or quadrantic deficit at either baseline or retest visits. In those migraineurs with visual aura, no relationship was found between side of visual aura and location of visual field deficit. Of the 12 MO subjects, 8 reported lateralized visual aura at their test visits 1 day after migraine. Of these eight MA subjects, five had a greater than expected decrease in sensitivity in one or both eyes for either TMP or SAP. In all these cases, the deficits were either superior or inferior arcuate scotomas that crossed the vertical midline of the visual field. Four of our MA subjects experienced aura symptoms in both hemifields. This is consistent with a population-based study of aura symptoms in which bihemifield aura symptoms were similarly found in approximately 30% of MA people tested. ²⁴  

Test–Retest Variability as a Function of Sensitivity for Individual Visual Field Locations

Figure 6 shows that many locations had significantly reduced sensitivity the day after a migraine, relative to baseline. Next, we explored whether locations with reduced sensitivity at baseline were associated with larger sensitivity losses after a migraine event. First, we determined the 5th and 95th percentiles of retest sensitivity distribution for each level of baseline sensitivity. For example, for all locations measured as 18 dB at baseline (pooled across eccentricity and subjects), we collated the sensitivity measured at retest and determined the 5th, 50th, and 95th empirical percentiles of this retest distribution. These data are shown in Figure 7a for SAP and 7b for TMP. Percentiles were determined only for baseline sensitivities occurring at least 20 times in the data set. The shaded area shows the range between the 5th and 95th percentiles for control subjects. 

Consistent with previous literature ^{25 26} Figure 7 shows an increase in test–retest variability with decreasing baseline sensitivity. This is apparent both for SAP and TMP. For both SAP and TMP, the lower 95th percentile of retest sensitivities was lower for migraineurs than for control subjects, across most of the range of baseline sensitivities. Inspection of Figure 7b reveals a large widening of retest sensitivities for migraineurs relative to control subjects for baseline TMP sensitivities of less than 21 dB. These locations had mild-moderate reductions of sensitivity measured at baseline, and often demonstrated deeper depression the day after a migraine. For example, 90% of locations in control eyes with sensitivities reported as 18 dB at baseline, retested with sensitivities of between 15 and 27 dB. In contrast, in migraineurs, 90% of locations with a sensitivity of 18 dB measured at baseline retested with sensitivities between 24 and 0 dB the day after a migraine. Figure 1 shows an example of a migraineur with this type of change. It should be noted that the migraineurs were not simply more variable than control subjects, as the 5th percentile of baseline versus retest performance was similar between groups indicating a similar number and degree of relatively improving locations. 

Performance of Migraineurs at 1 Week after Migraine

All migraineurs were tested 1 week after migraine. We were interested in determining whether sensitivity had returned to baseline in those individuals with deficits measured the day after a migraine. Box plots of the global indices measured at 1 week after migraine are shown in Figure 8 , along with those for 1 day after migraine for comparison. Group comparisons revealed no significant improvements in the global indices at 1 week relative to 1 day after migraine for either TMP or SAP (paired t-tests, all P > 0.05). There were, however, several individuals with localized deficits measured with SAP the day after a migraine that were not repeated a week later. For TMP, however, performance at 1 week was remarkably similar to that measured at 1 day after migraine in most subjects. Hence, there was a subgroup of migraineurs who demonstrated localized visual field deficits to flickering stimuli the day after a migraine that were not resolved 1 week later. 

Discussion

Consistent with previous reports, ^{1 2 3 4 5} this study demonstrates that visual field deficits are common in people with migraine at times between attacks. This study extends previous work by demonstrating that such visual field abnormalities are often more pronounced the day after a migraine, and, in many cases, are still present a week later. These visual field deficits are not a residual consequence of visual aura, as we found similar patterns of deficits in both aura and nonaura groups. Therefore, what are the possible causes and significance of these functional deficits? 

It is possible that some of the dysfunction measured the day after a migraine can be explained due to effects of fatigue or poor concentration after migraine, or alternately antimigraine medications. We assessed performance 1 day after the offset, rather than onset, of all symptoms to minimize these effects. Nevertheless, fatigue or medication effects would be expected to result in a generalized sensitivity loss across the field. We found a mild change in AD (approximately −0.5 dB) to the flickering stimuli, which may be explained by fatigue or aversion to the task. However, this deficit was not resolved a week after migraine and hence is not readily explained by these factors. Furthermore, the greatest changes in visual field performance after migraine consisted of localized deficits, which are not readily explicable by fatigue or aversion. 

The day after a migraine, a subgroup of the migraineurs demonstrated localized visual field deficits that were either absent, or present in a milder form, at their baseline test visits. Many of these deficits were unilateral and arcuate; indeed, no subject demonstrated a bilateral homonymous deficit. As the symptomatology of migraine is cortical, the appearance of precortical visual field deficits in individuals with migraine may seem counterintuitive. However, such deficits have been consistently reported by us ^{4 5 27} and others ^{1 3} in periods between migraine attacks. 

Although the mechanism underlying localized field deficits in our migraine group cannot be ascertained from this study, other studies suggest a localized vascular event, possibly at the level of the optic nerve head, as a plausible explanation. Migraine is considered to be a neurovascular condition, whereby neural events result in the alteration of blood flow causing pain and further nerve activation. ²⁸ Despite migraine not being a primary vascular condition, several research groups have found migraineurs to have poor peripheral vascular regulation at times between migraine and have proposed that this may increase the risk of NTG in these individuals. ^{12 29 30} Broadway and Drance ³¹ studied peripheral vascular flow in individuals with glaucoma and found that the presence of vasospasm and migraine was higher in people with focal ischemic type optic discs, than those with other types of glaucomatous cupping. There is also evidence that patients with glaucoma ³² and also those with migraine may have altered regulation of the potent vasoconstrictor endothelin. ^{33 34} These studies suggest altered perfusion, possibly at the optic nerve head, as a plausible explanation for the localized functional deficits measured in our study. 

Greater changes in visual field performance after migraine were identified with TMP than with SAP. Identifying different degrees of loss with TMP than SAP is consistent with our previous work. ^{4 27} Visual-function–specific perimetry (such as TMP and FDP, which preferentially assess the magnocellular visual pathway, and SWAP, which assesses the koniocellular visual pathway) have also been found to detect functional loss earlier than SAP in glaucoma. ^{35 36 37 38 39} It has been proposed that the larger, sparser neurons of the magnocellular and koniocellular pathways are either more susceptible to damage possibly due to the large axons having a metabolic disadvantage in adverse conditions, ^{40 41} or alternately that the absence of neural redundancy in these pathways makes early functional loss easier to detect. ⁴² As we have previously identified deficits consistent with both M and P pathway loss in people with migraine ²⁷ and some of migraine subjects in our present study demonstrated SAP deficits, it seems unlikely that the TMP deficits measured in the current study were due to selective magnocellular pathway involvement. 

In addition to a decrease in visual field performance after migraine, we found greater test–retest variability in our migraine group relative to control subjects. This more variable visual field performance, related in part to duration after migraine, should warrant exclusion of migraineurs from normative perimetric databases. Such exclusion, predicted from population estimates of migraine prevalence to be approximately 10% to 15% of the normative database, ¹⁴ may result in a tightening of the normative confidence limits of test–retest variability. As these limits are used in statistical analysis to classify visual field progression, it is possible that exclusion of the migraine group may enhance the ability to track visual field progression in other disorders. 

In this study, we tested young people with migraine who definitely did not have glaucoma. The longer-term significance, if any, of these visual field deficits is currently unknown, and longitudinal data would be needed to clarify this issue. We assume from our previous case study of a subject with migraine without aura that these deficits eventually resolve. ⁴ However for an interim period after migraine, there is a subgroup of clinical patients with repeatable visual field deficits, that in some instances masquerade as glaucomatous. The presence of a subgroup of clinical patients with greater than usual test–retest variability and periodic prolonged decreased sensitivity after migraine has implications for both differential diagnosis in a clinical setting, and clinical research studies using perimetry. 

 Supported in part by a Raine Medical Research Foundation Priming Grant (AMM). AMM is supported by National Health Medical Research Council (NHMRC) Australian Clinical Research Fellowship 139150.

 Submitted for publication June 30, 2003; revised November 10, 2003; accepted November 13, 2003.

 Disclosure: A.M. McKendrick, None; D.R. Badcock, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Visual fields measured for the left eye of a participant with migraine without aura (34-year-old woman) using both SAP (a, b) and TMP (c, d), measured at baseline (7 days after migraine) and the day after a migraine. Left to right: the sensitivities returned for each location (dB); a grayscale plot of performance relative to the age-normative database; Age Normal (AN) probability plot, which identifies locations with significant reductions in sensitivity compared with the normative database; an Hill of Vision (HoV) probability plot, which identifies locations with significant reductions in sensitivity compared with the patients own HoV. This subject is identified as subject 3 in Figures 2 and 6 . **P < 0.01 **P < 0.001.

Figure 2.

 

The duration after migraine at the baseline visit. Subjects numbered 1 through 10 are from the migraine without aura group, and subjects numbered 11 through 22 are from the migraine with aura group. The numbering of subjects is kept consistent with that of Figure 6 , with subjects being shown in ascending order of duration after migraine at baseline visit. Dotted horizontal line: 7 days after migraine.

Figure 3.

 

Box plots of the distribution of perimetric global indices for the migraine and control groups: AD for SAP (a) and TMP (c) and PD for SAP (b) and TMP (d). Shaded box shows the 25th, median, and 75th percentiles, and the whiskers the 10th and 90th percentiles. All outliers are shown. The left side of each panel shows baseline performance and the right side shows retest performance (1 day after migraine for migraineurs).

Figure 4.

 

Scatterplots comparing the global indices for SAP for the baseline visit and retest visit. For migraineurs, the retest visit 1 day after migraine is plotted. (a, b) AD index; (c, d) PD index. Control subjects are shown in (a) and (c). Migraineurs are shown in (b) and (d). (○) The five subjects with baseline visits conducted within 7 days after migraine; (□) the four subjects with the longest duration after migraine at baseline. Within these groups, individual subjects are numbered consistent with Figure 2 . Gray shaded area: subjects with worse performance at retest than at baseline.

Figure 5.

 

Scatterplots comparing the global indices for TMP for the baseline and retest visits. For migraineurs, the retest visit 1 day after migraine is plotted. Panel labels, symbols and subject numbers are as explained in Figure 4 .

Figure 6.

 

The number of visual field locations with a significant decrease (▪) or increase (□) in sensitivity measured the day after a migraine. Locations were considered to be worse or better on retest if the baseline–retest score was outside the location and procedure-specific two-sided 95% confidence limit of control baseline-retest performance. SAP (a, b) and TMP (c, d) in right and left eyes, respectively.

Figure 7.

 

Retest limits of performance on (a) SAP and (b) TMP in the migraine and control groups as a function of sensitivity at baseline. The 5th, 50th, and 95th percentiles of retest performance are shown. Shaded area: region between the 5th and 95th percentile of control retest performance.

Figure 8.

 

Box plots of perimetric global indices for the migraine subjects measured at 1 day and 1 week after migraine. AD (a, c) and PD (b, d) on SAP and TMP, respectively. Shaded box shows the 25th, median, and 75th percentiles, and the whiskers the 10th and 90th percentiles. All outliers are shown.