Relationship Between Tear Volume and Tear Meniscus Curvature

Objective  To investigate the relationship between tear volume and tear meniscuscurvature by means of the video meniscometer.

Methods  Eleven eyes of 11 patients with severe dry eye were studied (all female;mean ± SD age, 66.2 ± 7.7 years; 7 left eyes and 4 right eyes),each of whose puncta had been therapeutically occluded. Four instillationsof balanced salt solution were given at intervals of 1 minute in each studiedeye, with the concentration increasing in 5-µL steps from 5 to 20 µL.Before and after the instillation of balanced salt solution, tear meniscuschanges were recorded by video meniscometer and radius of the meniscus wascalculated from the printed images by means of the concave mirror formula.

Results  The mean radius of the meniscus increased linearly with increased dropvolume (r² = 0.65, P<.001), with mean ± SD radius values of 0.24 ± 0.08mm at baseline and 0.48 ± 0.13, 0.62 ± 0.13, 0.84 ± 0.26,and 1.00 ± 0.32 mm after separate instillations of 5, 10, 15, and 20µL of balanced salt solution, respectively. For each subject, a significantlydifferent slope defining the relationship between instilled volume and meniscusradius was seen (0.016-0.063 mm/µL; mean ± SD, 0.039 ±0.015 mm/µL), which was thought to depend on the difference in capacityof the fluid reservoir over the ocular surface.

Conclusions  This study confirmed that the volume of instilled eyedrops is linearlyrelated to the resulting radius of the tear meniscus, suggesting that thisradius is a useful measure in monitoring the tear volume. This is likely tohave implications both for dry eye diagnosis and for confirming the efficacyof punctal occlusion in this condition.

In human eyes, the total tear volume is distributed among 3 continuouscompartments: the culs-de-sac, menisci, and preocular film.¹ Themenisci act as a reservoir supplying fluid from which the preocular film isre-formed at each blink. They also accommodate the excess tears that accumulatewith reflex tearing, from lacrimal drainage obstruction, or after drop instillation.The menisci and tear film are held in place by interfacial forces.² Tear volume is positively correlated with lacrimalsecretory rate,³ and meniscus volume is reducedin tear-deficient dry eye.^4,5 Hence,knowledge of meniscus volume may be useful in the diagnosis of dry eye. Also,in dry eye, meniscus radius can be informative about the efficacy of punctalocclusion, either directly or by demonstrating impaired clearance of an instilleddrop.

Reflective meniscometry^5-7 isa noninvasive technique for the measurement of tear meniscus curvature. Themethod, which uses a specular reflection technique, projects a target ontothe concave, mirrorlike surface of the meniscus, and the size of the reflectedimage is used to calculate its radius of curvature. With certain assumptionsas to the sagittal profile of the meniscus, radius of curvature can be usedto estimate meniscus volume.^4,5

Recently, a video-based system (the video meniscometer)^5,7 hasbeen developed and used to demonstrate a significant relationship betweenthe meniscus radius and height.⁷ Both measureshave been used in the diagnosis of dry eye.^4,5,8-10 Thepresent study aims to assess the relationship between radius of tear meniscusand total tear volume.

In this study, 11 women with severe dry eyes were enrolled (mean ±SD age, 66.2 ± 7.7 years; range, 58-85 years). For each patient, oneeye was randomly selected for study (7 left eyes and 4 right eyes). The 11patients included 8 with Sjögren syndrome, diagnosed on the basis ofthe Fox criteria,¹¹ and 3 patients with non-Sjögrentear-deficient dry eye. All patients either had had upper and lower punctalplugs successfully inserted (8 eyes; Ready-Set M-size Punctum Plug; FCI Ophthalmics,Issy-les-Moulineaux, France), or had undergone upper and lower surgical punctalocclusion by cauterization or suture closure (3 eyes). All eyes showed theseverest form of dry eye, with a diminished Schirmer I test¹² value(mean ± SD, 1.5 ± 2.5 mm in 5 minutes) and, before occlusion,severe corneal fluorescein staining scores^13,14 (A3D3in 8 eyes, A3D2 in 2 eyes, and A2D3 in 1 eye, where A represents area andD, density, scored from 0 to 3 depending on the severity) and a total rosebengal staining score¹⁵ of 7.1 ± 2.2(mean ± SD) for the cornea and nasal and temporal bulbar conjunctiva.Before punctal occlusion, the patients had been treated with a combinationof preservative-free artificial tears containing 0.1% potassium chloride and0.4% sodium chloride (Softsantear; Santen Pharmaceutical Co, Ltd, Osaka, Japan)and hyaluronate sodium ophthalmic solution (HyaleinMini 0.1; Santen PharmaceuticalCo, Ltd) with little improvement. However, after occlusion therapy, fluoresceinstaining scores improved to A0D0 in all cases, and all subjects continuedtreatment with preservative-free artificial tears alone, 4 to 6 times daily.

This group of patients with severe dry eye was selected to provide amodel system in which the effects of instilled eyedrops on the radius of thetear meniscus could be isolated from the influence of reflex tearing and lacrimaldrainage. It was assumed that secretion was close to zero and that the artificialtears provided the major proportion of the tear reservoir in the conjunctivalsac, meniscus, and preocular film. On the day of the study, subjects wererequested not to use their artificial tear preparation until after the examination.

This research conformed to the tenets of the Declaration of Helsinkion the use of human subjects. Before the start of the study, procedures werefully explained to all subjects and informed consent was obtained.

In this study, we used video meniscometry as a noninvasive method toobserve and quantify tear meniscus curvature at the center of the lower eyelidmargin.^5,7 The system comprisesan illuminated target, CCD camera, video recorder, and video printer. Thehorizontally striped target is projected onto the center of the lower meniscus,which behaves like a cylindrical, concave mirror.

The line separation in the reflected image is used to calculate meniscuscurvature by means of the concave mirror formula.^5-7

The target consists of a series of black metal bars, 4 mm wide and 4mm apart, set in front of the objective lens and illuminated from behind.A half-silvered mirror, introduced into the optical pathway, permits coaxialviewing and image capture as well as back illumination of the target. Thereflected image appears as a series of black and white stripes (Figure 1). For the purpose of calculating the radius of the tearmeniscus, selected meniscus images were printed (print magnification, ×70). In normal subjects, in the absence of drop instillation and with thesubject blinking spontaneously, the reflective images of the target are highlystable over time.

To evaluate the relationship between radius of tear meniscus and instilledfluid volume, 4 instillations of balanced salt solution (BSS) were given atintervals of 1 minute into the selected eye of each subject in 5-µLsteps, increasing from 5 to 20 µL (ie, 5, 10, 15, and 20 µL).After each instillation, patients were allowed to blink spontaneously for30 seconds, after which tear meniscus behavior was recorded with the videomeniscometer. After the meniscus appearance was recorded, the radius was calculatedfrom the video printed images. The procedure for monitoring tear meniscusbehavior at the center of the lower eyelid margin in a typical subject's eyeis shown schematically in Figure 1.

The results were expressed as mean ± SD. Regression analysiswas used to estimate the straight-line relationship between the radius oftear meniscus and the accumulated volume of instilled eyedrops in each patient.Differences in slopes of the regression line between patients were analyzedby analysis of covariance with volume of eyedrops as a covariate. P≤.05 was considered to be significant.

Before drop instillation, the mean baseline meniscus radius was 0.24± 0.08 mm. After the cumulative instillation of 5, 10, 15, and 20 µLof BSS, the mean meniscus radius was calculated to be 0.48 ± 0.13,0.62 ± 0.13, 0.84 ± 0.26, and 1.00 ± 0.32 mm, respectively.A representative example of the change in meniscus image after successiveinstillation of BSS is shown in Figure 2.

The mean radius increased linearly with total volume instilled (r² = 0.65, P<.001; y = 0.038x + 0.256, where y = radius [in millimeters] and x =total volume of instilled BSS [in microliters]; Figure 3). The significant regression between radius and total volumeinstilled was found for each subject individually (0.039 ± 0.015 mm/µL[0.016-0.063 mm/µL]; Table 1 and Figure 4). Also, analysis of covariance showedthat there was a significant interaction between volume of eyedrops and patients(F = 12.33, P<.001; Table 2), implying that slopes of regression line between patientswere significantly different.

One view of tear-deficient dry eye is that it is the result of an autoimmuneinflammatory process that destroys lacrimal tissue or function and resultsin reduced tear secretion.¹⁶ The autoimmuneprocess can result, directly or indirectly, in inflammatory events at theocular surface, particularly the conjunctiva, and lead to tear hyperosmolaritydue to evaporation from a reduced aqueous pool. Hyperosmolarity itself leadsto surface inflammation.^16-19 Intear-deficient dry eye, a reduced tear flow leads to a reduced tear volume.It has been suggested that this in turn leads to a thinning of the preoculartear film^2,20 and that this isthe major factor leading to hyperosmolarity.^17-19,21 Amathematical relationship between meniscus curvature and tear film thicknesshas been proposed,^20,22 and forthis reason alone, we would expect meniscus curvature to be of diagnosticvalue in dry eye. In this study, we attempted to show in addition that meniscuscurvature, which relates to its volume, is an index of total tear volume.This again would support its potential value in diagnosis of dry eye.

For the purposes of discussion, the term total tearvolume will be used to mean the aggregate volume of tears distributedbetween the culs-de-sac, menisci, and preocular tear film.¹ Meniscusvolume is chiefly regulated by the balance between lacrimal secretion andtear drainage, with evaporation playing a minor role. Evaporation is, however,of increasing importance in dry eye and becomes the dominant feature of evaporativedry eye.^23-26

The present study aimed to assess the relationship between tear meniscusradius and total tear volume under conditions in which total tear volume couldbe increased incrementally. This was achieved by studying a group of aqueous-deficientpatients with severe dry eye, whose tear secretion was minimal (as indicatedby the Schirmer I values before the punctal occlusion), and lacrimal drainagewas completely obstructed by punctal occlusion. In this situation it was reasonablyassumed that the total volume of fluid bathing the ocular surface was determinedchiefly by the experimental instillation of BSS. Evaporation was assumed tohave a negligible effect on tear volume during the short time course of thestudy. Relative to the total volume of BSS instilled, the evaporative lossduring the course of the study, calculated according to a previous study²⁷ for a dry eye population, would be about 0.23 µLper eye, which would be unlikely to have a major effect on the outcome ofthe study. Since all patients exhibited a visible meniscus at baseline, itmust be assumed that an amount of fluid was present at the start of the experiments,reflecting the persistence of residual secretions or of tear substitutes instilledpreviously.

As mentioned in the "Results" section, a linear relationship was shownbetween the radius of the lower central tear meniscus and the cumulative volumeof instilled fluid, both in the total group (Figure 3) and in individual subjects (Figure 4, Table 1),indicating that the radius of curvature of the meniscus reflects the totalvolume of tears distributed between the tear compartments at any one time.These findings allow the general conclusion to be drawn that meniscus radiusvalues derived from measurements over a small region of the lower centraltear meniscus are probably generalizable to the whole meniscus reservoir,and that the radius is not significantly influenced by any change in surfaceactive properties of the tear fluid²⁸ due todilution.

Extrapolation of the group mean regression line to the x-axis in Figure 3, where the radius is notionallyzero, may be interpreted to reflect the baseline total volume of tear fluiddistributed in sac and film. This value was 6.7 µL for the group. Ifwe take the reported average radius value in normal subjects to be 0.365 mm,using the same video meniscometric technique,⁵ thenthe relationship plotted for the group in Figure 3 shows that this radius would correspond to an additionalvolume of 2.9 µL above the baseline value for the dry eye group. Ifthe baseline total dry eye volume (6.7 µL) is added to this additionalvolume (2.9 µL), the sum (9.6 µL) is close to the value reportedpreviously¹ for the tear volume in the normaleye. We recognize that this is a crude estimate, but we submit it here toshow how such a calculation could be used in the future with more refinedtechniques.

Figure 4 demonstrates importantindividual differences between the patients with dry eye in this study. Itshould be noted that radius values differed between subjects and that forsome subjects these values fell within the reference range. This raises thepossibility that these subjects did not comply with instructions not to instilltear substitutes on the day of study or that artificial tear preparation instilledbefore the day of the experiment remained on the eye because of the completeocclusion of the puncta. Also, not only do basal radius values differ betweenpatients, but the slopes of the regression lines also significantly differ(0.039 ± 0.015 mm/µL [0.016-0.063 mm/µL]; Table 2), implying that the instilled volumes of BSS were accommodatedinto compartments of differing properties. The most likely factor that mightinfluence these slopes would be cul-de-sac size and physical properties, withsurface tension and tear viscosity also playing a small role. Since the tearsacs are not rigid compartments, it might be assumed that elastic factorswould influence the response to added volume and that this might be modifiedby disease.

The differences in individual slopes of the radius-volume curves drawattention to the likelihood that the same meniscus radii in 2 different individualsmay not imply the same total tear volume, probably reflecting differencesin sac capacity. On the other hand, in a given individual, changes in radiuswill directly relate to changes in total tear volume. This is of importancein relation to studies of tear clearance or in clinical trials, particularlyof lacrimal secretogogues.

In summary, the present study indicates that meniscus curvature is anindex of total tear volume, while mathematical considerations suggest thatit predicts tear film thickness.^21,22 Theseobservations support the theoretical value of meniscus curvature measurementand of video meniscometry in the diagnosis of dry eye. We will show elsewherethat the same technology can be used to confirm the efficacy of punctal occlusionin this condition.

Correspondence: Norihiko Yokoi, MD, PhD, Department of Ophthalmology,Kyoto Prefectural University of Medicine, 465 Kajii-cho, Hirokoji-agaru, Kawaramachi-dori,Kamigyo-ku, Kyoto 6020841, Japan (nyokoi@ophth.kpu-m.ac.jp).

Submitted for publication July 17, 2003; final revision received November17, 2003; accepted April 6, 2004.

This study was supported in part by grants-in-aid for scientific researchfrom the Japanese Ministry of Health, Labor, and Welfare and the JapaneseMinistry of Education, Culture, Sports, Science, and Technology (Tokyo) andby research grants from the Kyoto Foundation for the Promotion of MedicalScience, the Intramural Research Fund of Kyoto Prefectural University of Medicine(Kyoto, Japan).

This study was presented in part at the annual meeting of the Associationfor Research in Vision and Ophthalmology; April 30, 2000; Ft Lauderdale, Fla.

We thank Tatsuo Terashima and Kousei Terasaki (Biostatistics Department,Santen Pharmaceutical Co Ltd, Osaka, Japan) for statistical assistance.