In 1964, Taylor1 examined a young woman with moderate myopia sent by her driving instructor for presumed defective lateral vision. She wore glasses with a “fashionable” frame described as “heavy,” with “very wide sides” and which, in Taylor’s view, caused severe visual field constriction. Different studies have investigated the peripheral visual field in healthy volunteers with different frame styles.2–4 However, the methodology used in these studies did not allow quantification of visual field surface areas, making it impossible to state how severe the constriction was for a given frame style.
In California, it has been unlawful since September 18, 1959, to drive a motor vehicle with glasses having a temple width of one half-inch or more (Section 23120 of the Motor Vehicle Code of the state of California to be found at http://www.dmv.ca.gov/pubs/vctop/d11/vc23120.htm). But overall in the literature, the liability of wide temples in temporal visual field constriction has probably been understated. Thus, thick frames were considered to cause more apparent temporal field constriction than wide temples3 and thick frames with wide temples were reported to cause temporal scotomata, but only the extreme temporal field was involved.4 Even more strikingly, in one report,2 the frame with the widest (17.4 mm) temples constricted the binocular field below the 140° minimum safety requirement in only 9 of 75 subjects. This safety requirement was issued in 1964 by the Federal Motor Carrier Safety Administration for commercial motor vehicles drivers (see http://www.mrb.fmcsa.dot.gov/documents/PPP/Vision_and_Commercial_Motor_Vehicle_Driver_Safety.pdf).
These reports imply that wide temples were not harmful to the temporal visual field. While seemingly evident, Taylor’s view that thick frames with wide temples cause severe visual field constriction1 has not yet been convincingly substantiated. Understanding the effects of eyewear styles on vision and safety could be beneficial as fashion trends change, and this type of frame is currently popular.
Using the Goldmann perimeter, we recently reported a dramatic visual field increase through eye motion, especially in the temporal quadrant.5 The method we used allowed calculation of visual field surface area in the primary position of gaze, then allowing eye motion. A headband prototype fitted with a line laser allowed controlled head rotation, necessary to test temporal points falling out of the cupola extent. Using the same method, we set out to compare two sunglass styles: one with thin temples and a thin frame (referred to as “thin sunglasses”) and one with very wide temples and a thick frame (referred to as “thick sunglasses”). Our aim was to quantify the influence of these types of temples/frames on the visual field using the primary position of gaze but also with allowed eye motion. The latter evaluation was deemed to be closest to real-life conditions, thus particularly useful in proving how harmful wide temples were to the temporal visual field.
Visual Field Measurement
The method used to test the visual field has already been described in a previous report.5 Briefly, the visual field was tested with the Goldmann perimeter using the V4 (biggest and brightest) test object, from seen to unseen, in 15 healthy volunteers. The volunteers (fellows and nurses) have been fully identified in a previous publication.5 They were all personally known to the authors. The guidelines of the Declaration of Helsinki were followed, and informed consent was obtained in each case.
A headband prototype fitted with a magnetic line-laser level having two orthogonal vial levels (GPLL5, Bosch) was used to control head position. The visual field was tested, first in the primary position of gaze, then allowing eye motion. In both cases, some temporal points falling outside the cupola were retested after a controlled nasal head pure rotation (using the headband/line-laser level). Using this protocol, we tested the “base visual field” (BVF), i.e., visual field in primary position of gaze. We then tested the “eye motion visual field” (EMVF), i.e., visual field with allowed eye motion. Both BVF and EMVF were tested without glasses and then with two different sunglasses (Fig. 1). These sunglasses are manufactured by the Maui Jim Corporation (Hawaii): reference “117-02 / Akoni” (“thin sunglasses”) and reference “203-02 / Lehua” (“thick sunglasses”). Both of these sunglasses have exactly the same type of lens (lens color: “Neutral grey”; polarized lens type: “PolarizedPlus®2”). The largest dimension (width and height) of the glasses (measured respectively in a horizontal and a vertical plane), taking the forward bowed and curved glasses shape into account, was 62 mm × 44 mm for the “thin sunglasses” and 62 mm × 43 mm for the “thick sunglasses.” Both sunglasses were bowed forward in a comparable fashion. The two sunglasses differed mainly in that the “thin sunglasses” had a thin metal frame with thin (3.5 mm) and high-set temples, whereas the “thick sunglasses” had a thick plastic frame with wide (26.5 mm) temples (Fig. 1).
In the case of momentary disappearance of the V4 test object (hidden by the frame or a temple), reported only with the “thick sunglasses,” only the most peripheral point perceived beyond the frame or temple was taken into account.
A “Chromophare D 300” surgical lighting system (Bertchold) fitted with a 12-V/50-W halogen light was used for the glare test. The surgical lighting system was positioned at 75 cm from the subject’s eye(s), and the light beam was pointed at the eye(s) at different incidence angles: frontal, lateral, inferior (30° from vertical), and superior (30° from vertical). We aimed at measuring discomfort glare, defined as that caused by excessively intense illumination.6 For each incidence, the subject was asked which type of sunglasses offered the least amount of annoyance from the illumination from the surgical lighting system.
The frame sequence was randomized for both visual field and glare tests.
As described previously,5 spherical coordinates system (ρ; θ; ϕ) was used to calculate the surfaces (cm2) occupied by BVF and EMVF on a virtual cupola whose extent would be large enough to include all points and whose radius would be 30 cm, as with the Goldmann cupola. The surface area of a sphere of radius « r » equals 4πr2. The Goldmann cupola whose radius is 30 cm therefore has a surface area of 11,310 cm2. A sphere being 41,253 deg2, a surface area of 1000 cm2 projected on the Goldmann cupola, is equivalent to about 3647.5 deg2.
For some analyses, the visual field was split into four quadrants. The meridian under investigation being defined as α, quadrants were defined as follows: superior quadrant for α ∈ [45°; 135°], nasal quadrant for α ∈ [135°; 225°] for the right eye and α ∈ [315°; 45°] otherwise, inferior quadrant for α ∈ [225°; 315°], and temporal quadrant for α ∈ [315°; 45°] for the right eye and α ∈ [135°; 225°] otherwise.
After having verified the normality assumption, surface areas were presented as the mean (cm2) and standard deviation. The type I error of every statistical test was fixed at 5%. The reproducibility error of our method was estimated to be 14%.5 Visual field surface area differences larger than the 14% reproducibility error and having p < 0.05 were considered significant. For the discomfort glare test, p values were calculated using an exact binomial test, comparing the frequency of “less glare with ‘thin sunglasses’” and “less glare with ‘thick sunglasses’” to a 50-50 distribution, for the related light incidence.
Quantitative analysis and comparison of BVF and EMVF surface areas without sunglasses, with the “thin sunglasses” or with the “thick sunglasses,” are displayed in Tables 1 and 2, respectively.
There were no significant differences when comparing BVF or EMVF overall or quadrant surface area variations using the “thin sunglasses” versus no glasses. In contrast, a 22% EMVF decrease (p < 0.001) occurred when comparing monocular surface areas using the “thick sunglasses” versus no glasses. For BVF, this decrease was statistically significant in the nasal quadrant (−21%; p < 0.001) and superior quadrant (−20%; p < 0.001). For EMVF, the decrease was statistically significant in the nasal quadrant (−25%; p < 0.001) and temporal quadrant (−33%; p < 0.001).
Fig. 2 shows combined BVF and EMVF plots of three subjects without glasses, with the “thin sunglasses” and with the “thick sunglasses.” These plots were chosen because they were a representative range of EMVF constriction patterns found using the sunglasses in the 15 healthy volunteers. In Fig. 2A (subject 15), the top plot (no glasses) and the middle plot (“thin sunglasses”) closely resemble each other. The bottom plot (“thick sunglasses”) shows a comparable BVF aspect to without glasses or with the thin sunglasses. In contrast, the bottom plot shows a severe EMVF constriction. This constriction involves the temporal quadrant, where some inferotemporal visual field is spared below the wide “thick sunglasses” temple. This is the first of three EMVF constriction patterns we identified with the “thick sunglasses.” This first EMVF constriction pattern with the “thick sunglasses” was also found in the plots of both eyes of subjects 1, 5, 8, 10, 13, and 14 (data not shown). In Fig. 2B (subject 11), the top plot (no glasses) and the middle plot (“thin sunglasses”) closely resemble each other. The bottom plot (“thick sunglasses”) shows some BVF constriction, especially in the temporal quadrant, and a severe EMVF constriction, especially in the temporal quadrant. This EMVF constriction involves some inferotemporal and superotemporal visual field spared below and above the wide “thick sunglasses” temple, respectively. This is the second of three EMVF constriction patterns with the “thick sunglasses.” This pattern was also found in the plots of the left eye of subject 2 and in the plots of both eyes of subjects 6, 7, and 12 (data not shown). In Fig. 2C (subject 9), a moderate BVF and EMVF constriction between the top plot (no glasses) and the middle plot (“thin sunglasses”) occurs. The bottom plot (“thick sunglasses”) shows a dramatic concentric constriction for both BVF and EMVF. This third EMVF constriction pattern with the “thick sunglasses” was also found in the plots of both eyes of subjects 3 and 4 (data not shown). This concentric constriction may be explained by the fact that, in subjects 3, 4, and 9, the thick plastic frame of the “thick sunglasses” was very close-fitting. All eyes followed one of these EMVF constriction patterns except for the right eye of subject 2, in which some superotemporal field (“above the temple”) was spared.
The discomfort glare test was aimed at assessing which type of sunglasses offered the best protection from the illumination from the surgical lighting system. The results of this test are displayed in Table 3. All subjects reported less lateral glare with the “thick sunglasses” (p < 0.001). Overall, the “thick sunglasses” seemed to offer more protection for every light incidence but the frontal.
In past studies, regardless of the method used, only the base visual field (BVF), i.e., the visual field in the primary position of gaze, was used to assess the effect of thick frame styles with wide temples on the visual field, and no visual field surface area quantification was undertaken. These studies failed to convincingly prove how harmful such frame styles were to the visual field.2–4 In contrast, our method5 allows a rational comparison of two different sunglass frame styles through visual field surface area quantification of BVF and EMVF, i.e., visual field with allowed eye motion. The latter is characterized by a dramatic visual field surface area increase in the temporal quadrant.5 For this reason, we believed that EMVF was more likely than BVF to demonstrate wide temple–related temporal field constriction. In this regard, it should be noted that, in 1966, Smith and Weale7 showed that, in subjects turning their head to look back with their eye in abduction, frames with wide temples and thick frames reduced the “central field” to a marked degree. However, testing “central field” on an abducted eye amounts to testing part of the mid peripheral EMVF. It is therefore interesting that, in aiming to conduct a study on “central field,” Smith and Weale inadvertently showed that EMVF was a factor by which the harm done by wide temples and thick frames was clearly demonstrated.
The present study shows that sunglasses with thick frames and wide temples severely constrict the visual field. Using the V4 test object of the Goldmann perimeter, no marked differences between BVF and EMVF tested without sunglasses or with the “thin sunglasses” were demonstrated. This suggests that, in daylight conditions the type of tinted glasses used for the study does not result in severe visual field constriction. The thickest element of the “thin sunglasses” is the anterior part of the temples (3.5 mm). The negligible impact of the “thin sunglasses” may be explained by the fact that the remaining part of the temples and the frame are much thinner,2 probably thinner than most pupil diameters,4 and also that the temple attachment is high2 while the maximum temporal eccentricity of EMVF is, on average, on the horizontal meridian.5 In contrast, the “thick sunglasses” result in a 22% EMVF surface area decrease (p < 0.001). The “thick sunglasses” effect on BVF surface area is significant in the superior (−20%; p < 0.001) and nasal quadrants (−21%; p < 0.001) and suggests that the thick plastic frame plays a part in visual field constriction. In subjects in whom, owing to anatomical reasons, the sunglasses are very close fitting, a concentric visual field contraction may occur (3 of 15 subjects) for both BVF and EMVF (Fig. 2C). Our method failed to detect a significant BVF surface area constriction in the temporal quadrant, the 12% constriction being inferior to our reproducibility error. In contrast, the effect of the wide temples on EMVF surface area in the temporal quadrant was markedly elevated (−33%; p < 0.001). Contrary to BVF, EMVF analysis therefore clearly shows that the wide temples of the “thick sunglasses” act on the temporal visual field as “blinders act on a horse.”2 The average EMVF surface area in the temporal quadrant with the “thick sunglasses” (1276 cm2) is inferior to the BVF surface area in the temporal quadrant with no glasses (1323 cm2). In summary, wearing “thick sunglasses” greatly minimizes our ability to explore lateral space by resulting in a loss of most, if not all, of the additional temporal visual field gained through eye motion.
The better protection from lateral glare (p < 0.001) offered by the “thick sunglasses” obviously relates to their wide, solid temples. By blocking peripheral light, these wide temples may also act as a protection from posterolateral light focusing by the cornea to the nasal limbus, a phenomenon most probably implicated in pterygium pathogenesis.8–10 For the same reasons, it may also prevent photokeratitis, a condition already described in Antarctica for different sunglasses styles with little lateral protection.11 The thick plastic frame of “thick sunglasses” may at least account for the better protection from inferior and superior glare (p = 0.039). These advantages of the “thick sunglasses” over the “thin sunglasses” are offset by the negligible incidence of the latter on the visual field.
Some frames available with prescription lenses have been manufactured with wide temples constructed with an artistic design similar to wrought-iron work. This design interferes less with lateral vision and may therefore reduce the problem of temporal field restriction resulting from wide temples. Sunglasses fitted with wide temples having a central area made of transparent dark-tinted sun lens material have already been manufactured and are likely to offer protection from posterolateral light-related eye damages and interfere less with lateral visual field than sunglasses with solid, opaque wide temples. Similarly, darkly tinted polycarbonate safety frames with very wide temples have also been manufactured. Such designs cause little interference with the temporal visual field and lessen the safety hazard.
Using a driving simulator and on-road driving conditions to study individuals with peripheral visual field defects, Coeckelbergh et al.12 showed that individuals who made more head movements and started scanning eye movements early were more likely to pass an on-road driving test. Such eye and head movements are likely an integrated part of physiologic visual field exploration in normal individuals.5 In our study, unlike the “thin sunglasses,” the “thick sunglasses” severely constrict the temporal visual field and may therefore hamper the eye movements related to lateral space exploration. The present study therefore substantiates the previously held view1–3,7 that such thick frame styles with wide temples present a danger to driving a motor vehicle. It should be mentioned that in California, since September 18, 1959, it has been forbidden to drive “a motor vehicle while wearing glasses having a temple width of one-half inch or more if any part of such temple extends below the horizontal center of the lens so as to interfere with lateral vision.” (Motor Vehicle Code, Section 23120, to be found at http://www.dmv.ca.gov/pubs/vctop/d11/vc23120.htm).
Department of Ophthalmology
Caen University Hospital
Avenue de la Côte de Nacre
The authors thank Miss Marie-Hélène Drouet (orthoptist) who carried out all visual field tests.
The authors report no conflict of interest.
The sunglasses used in this study were kindly lent to the authors by the Maui Jim Corporation (Hawaii).
This article has been presented as a poster (F119) at the 2012 European Association for Vision and Eye Research congress in Nice, France.
Received April 29, 2013; accepted June 23, 2013.