Lacrimal and meibomian gland evaluation in dry eye disease: A mini-review

Abstract

Lacrimal and meibomian glands contribute to the aqueous and lipid components of tear film, respectively. Their evaluation remains central to diagnosing and managing dry eye disease (DED). The review discusses the differences and reliability of various diagnostic tests and commercially available devices used for DED diagnosis. Slit-lamp-based techniques are direct palpebral lobe and tear flow assessment, Schirmer test, meibum quality and expressibility, and evaluation of tear meniscus height. Non-invasive tear meniscus height (TMH), tear break-up time (TBUT), lipid layer thickness (LLT), and meibography are machine-based diagnostic tests. The structure–function correlation of the tear-producing glands gives more comprehensive details than either information alone. Many devices are available in the market, which make DED diagnosis an easy feat, but the tests should be interpreted keeping in mind the intra-observer and inter-observer repeatability. Also, the tear film displays a huge variability as per the environmental conditions and impact of blinking. Hence, the examiner should be well versed with the techniques and repeat the test two to three times to obtain an average reading, which is more reliable. The recommended sequence of tests for diagnosing DED is a dry eye questionnaire, TMH, LLT, NIBUT (FBUT if non-invasive test is unavailable but should be performed after osmolarity), tear osmolarity, meibography, and ocular surface staining. Invasive tests such as Schirmer should be performed after the non-invasive tear film diagnostic testing.

Tear film in human eyes contains complex arrangement of lipid, aqueous, and mucin components. Lipid, aqueous, and mucin components are contributed by meibomian, lacrimal glands, and conjunctival goblet cells, respectively. The lacrimal gland is a compound tubuloacinar gland consisting of acini and an inter-connecting ductal system comprising intra-lobular, inter-lobular, and terminal secretory ducts varying in number from 4 to 7.[1-3] The lacrimal gland has two lobes: a larger orbital lobe and a smaller palpebral lobe divided by levator aponeurosis [Fig. 1]. The palpebral lobe can be seen on slit lamp examination, whereas the orbital lobe can be visualized using imaging modalities such as computed tomography or magnetic resonance imaging.[3] In patients with aqueous deficiency dry eye disease (ADDE), the lacrimal gland is affected either primarily like Sjogren’s syndrome or secondarily to ocular surface cicatrization, radiation, or secretomotor facial nerve palsy.[4-6] Meibomian glands are modified sebaceous glands located within the tarsal plate of upper and lower eyelids.[7] They have a large central duct where acini empty themselves into collecting ducts in a holocrine manner and central duct secretes meibum. Meibomian glands vary in number, and the maximum number seen on meibography in a large Indian cohort is 30 in the upper eyelid and 28 in the lower eyelid (unpublished data). Dry eye disease (DED) has two main sub-types, ADDE) and meibomian gland dysfunction (MGD), and a mixed form. ADDE occurs secondarily to lacrimal gland damage, and MGD occurs because of meibomian gland pathology.[8,9] Thus, examination of the lacrimal and meibomian glands is essential for diagnosing DED. This review discusses the various techniques of evaluating the lacrimal and meibomian glands based on tests available in our clinical practice.

Figure 1.

Schematic shows the location of lacrimal and meibomian glands in the eye. Also, the three components of the tear film, superficial lipid, middle aqueous (blue), and mucin layer (green), can be seen

Examination Techniques for Lacrimal Gland Function

Schirmer test

In 1903, Otto Schirmer introduced a test for measuring tear production using filter paper strips of 5 * 35 mm and placed them in the conjunctival sac for 5 min while subjects usually blink.[10] Today, commercially available Whatman 41 filter paper strip is used for Schirmer test, and it has markings for up to 35 mm. The proximal part of the strip, which sits in the conjunctival-cul-de-sac, is rounded, devoid of markings, and should not be touched with fingers. Both eyes are tested simultaneously, and at the end of 5 min, the moistened segment is measured. The transitional zone (junction of wet and unwet) should be seen carefully in good light immediately after removal, and if difficult, trans-illumination can be used. The strip should not touch the cornea to avoid reflex tear activity as the cornea is far more sensitive to touch than the eyelid margin or conjunctiva. Care should be taken, if possible, not to touch the lid margin with the strip while it is within the conjunctival cul-de-sac.[11] Initially, the Schirmer test was performed with open eyes, but measurements show less variation when obtained with closed eyes.[12] Historically, Schirmer I is measured without anesthesia and open eye, Schirmer II measures reflex stimulated tear secretion, that is, after nasal stimulation, and Schirmer III measures reflex-stimulated secretion by looking into the sun. In today’s world, it is best to define the Schirmer test with or without anesthesia and with open or closed eyes to avoid confusion. Measuring Schirmer values under anesthesia gives 60% of the unanesthetized value.[13] Instilling paracaine reduces Schirmer values but does not abolish it to zero, thus opening the debate whether the lacrimal gland is a basal secretor.[14] Norn suggested not to use anesthesia in Schirmer’s test as reflex tearing is unavoidable.

The cut-off value for the Schirmer test in diagnosing DED is 10 mm for 5 min.[8] Schirmer took 15 mm as the cut-off, Sjogren took 5 mm, and Jones took 10 mm, but Bjisterveld found 5.5 mm as the cut-off in 550 normal and 43 subjects with DED.[11,15] The sensitivity and specificity of the Schirmer test are 83% and 85%, respectively, using 5.5 mm as the cut-off value.[15] Other investigators have found sensitivity to be 10% and a high coefficient of variation ranging from 60–100%.[11,15] Nevertheless, the Schirmer test is still the most commonly employed test in dry eye evaluation and is easy to perform. TFOS-DEWS II has taken a value of less than 10 mm as one of the DED diagnostic criteria. Values less than 5 mm are definitely pathological, values more than 10 mm are normal, and values between 5 to 10 mm should be interpreted carefully.

Tear meniscus assessment

Meniscometry is the study of tear meniscus, which is a reservoir to supply tears to the precorneal tear film. Various parameters such as tear meniscus height (TMH), tear meniscus area (TMA), tear meniscus curvature (TMR), and tear meniscus depth have been studied.[16-31] There are slit-lamp-based meniscometry techniques and non-invasive optical coherence tomography (OCT)-based techniques. Low coherence interferometry and dim illumination of OCT reduce reflex tearing, which can erroneously raise TMH values when measured using a slit lamp.[19] TMH measured using OCT is more repeatable and reproducible compared to Schirmer test, which is invasive and induces reflex tearing. Although there is a poor agreement between time-domain and spectral-domain OCT (SD-OCT) for calculating TMH, individually, these machines have shown a good correlation with symptom scores and Schirmer values, although SD-OCT gives better resolution and reliability.[16,19,20,22,23] The reported normal mean TMH values are 0.29 mm (SD, 0.13 mm, slit lamp), 0.27 to 0.29 mm (SD, 0.05 to 0.12 mm, keratograph), and 0.19 to 0.34 mm (SD,0.02 to 0.15 mm, SD-OCT).[16-25,28] The diagnostic accuracy of SD-OCT is more in eyes with markedly reduced tear secretion like in Sjogren’s syndrome (91%; cut-off <0.141 mm) and poor for evaporative DED (45%; cut-off <0.256 mm) where TMH values can be more or less than normal.[24,25] For non-Sjogren’s ADDE, the diagnostic accuracy was 70% at a cut-off value of less than 0.248 mm. Patients with TMH values less than 0.21 mm, obtained using RTVue OCT, were at risk for corneal epithelial disease, and more than mean normal 0.345 mm TMH value in MGD patients correlated with corneal epithelial disease.[25] Hence, TMH values can be taken as a guide for patient management other than diagnosing DED. Another modality is swept-source OCT (SS-OCT), which gives more depth and resolution than the above two and can also give tear meniscus volume values.[26] The cut-off values for labeling DED using SS-OCT TMH, TMA, and TMV are 191 μm, 12,360 μm², and 0.0473 mm, respectively.[26] TMA provides 2D view of tear meniscus compared to TMH and has been found a little more sensitive in diagnosing DED.[24] The TMH values obtained with OCT machines depend on palpebral aperture width, lower eyelid laxity, conjunctivochalasis, and blink.[16] TMH values obtained using either OCT or a keratograph display no significant differences; hence, a keratograph can be used as a single machine for measuring TMH along with other parameters [Fig. 2].[30] Many groups have shown good correlation of TMH with Schirmer values, corneal staining scores, and TBUT.[29] However, the cut-off values of TMH in deciding the medical management and differentiating between mild ADDE (not so reduced Schirmer value) and EDE would be useful in clinical scenarios. Patients with Sjogren’s syndrome have low Schirmer values and are expected to have low TMH values, but how non-invasive TMH assessment impacts the disease management and its progression would be vital for clinical translation.

Figure 2.

Tear meniscus height obtained using Oculus keratograph 5M and measured manually in nasal, central, and temporal quadrants

Direct examination of the lacrimal gland

The palpebral lobe of the lacrimal gland can be examined on a slit lamp for morphological changes in DED patients. Also, the secretory ductules of the orbital lobe pass via the palpebral lobe; hence, the tear secretion, when assessed from the palpebral lobe, actually represents the whole lacrimal gland activity. Singh et al. reported the morphological changes in terms of size (exposed area of the gland), shape (convex or flat), and overlying conjunctival appearance (presence of engorged vessels, whitish areas of subepithelial scarring, or symblepharon).[3] The normal palpebral lobe has a convex shape, pinkish-colored with overlying normal conjunctiva, situated in the superotemporal fornix and extending up to or beyond the lateral canthus [Fig. 3]. In DED patients, the lobe becomes smaller in size with the flattening of contour and a change in the overlying conjunctiva’s appearance [Fig. 3]. The overlying conjunctiva may appear reddish with inflammation or have whitish areas of subepithelial scarring. Between SS and SJS, SJS patients have a higher proportion of lobes with a flat contour and subepithelial scarring with or without symblepharon.

Figure 3.

Palpebral lobe of lacrimal gland in normal and DED patients. a & e, Right palpebral lobe of a young healthy individual (Schirmer 27 mm) shows four secretory ductules (marked with an arrow) having a tear flow rate of 1.21 μl/min. b & f, Normal tear flow seen in a patient with MGD with three secretory ductules. c & g, Reduction in tear flow (0.23 μl/min) and secretory ductules in a patient with Sjogren’s syndrome. d & h, A Stevens–Johnson syndrome patient (Schirmer 0 mm) has flattening of the contour, whitish conjunctival scarring, and no secretory opening

Direct assessment of tear secretion (DATS) from the palpebral lobe has shown significant differences between MGD, SS-ADDE, and dry eye in cicatrizing conjunctivitis (CC). DATS was first published in 1986, where only the subjective description of the test was available.[31] Kim et al. quantified the tear flow from the palpebral lobes where reduction was observed in DED patients compared to normal.[32] The technique was further refined by Singh et al., where etiology-based differences in tear flow among various sub-types of DED were reported [Fig. 3].[33] A dry 2% sodium fluorescein ophthalmic strip is touched onto the exposed palpebral lobe. The tear secretion is visualized as fluorescein washout areas at the ductular openings. A video camera-attached slit lamp microscopy system is used, and then an image snapshot is captured at the end of 1 second after fluorescein touch. The area of each ductular opening is calculated in mm² using the Image J freehand tool. This area (in mm₂/sec) is then multiplied by 10 microns (assumed tear film thickness in that region) and 60 seconds to give the tear flow rate in microliters per minute. This is for calculating the precise volume of the research purpose, and it is zero in CC group (0 to 3) [P < 0.000001]. The median tear flow rate per lobe in CC (0.00 μl/min) and SS (0.21 μl/min) was significantly lesser than those of normal lobes (1.05 μl/min) and EDE (0.99 μl/min) eyes.[33] Within ADDE, the tear flow rate was significantly reduced in CC eyes compared to SS. A simple visual assessment of secretions from the lobe and the appearance of the lobe can give practitioners many clues about gland involvement in DED. Also, the time lag between applying the fluorescein strip and the first appearance of secretory activity tells about neurosensory lag (in seconds). There was an increase in time lag in ADDE patients compared to MGD. The correlation between Schirmer values and DATS was linear for Sjogren’s syndrome and SJS but nonlinear in normal and evaporative DED.[33] The tear flow rate was linear till 20 mm Schirmer values in normal and after that remained the same despite an increase in Schirmer values. Schirmer test measures volume of tears in conjunctival cul-de-sac, which is influenced by many factors such as blinking, tear evaporation, the lipid content of tear film, and eye drops. Hence, DATS seems more specific for evaluating the lacrimal gland function. Similar fluorescein-based test can be used for examining labial salivary glands in diagnosing SS patients and evaluate transplanted minor salivary glands.[34,35]

Examination techniques for meibomian gland function

Meibomian glands are modified sebaceous glands that secrete a complex mixture of lipids and also some proteins into the eyelid margin, which spreads over the ocular surface [Fig. 1]. MGD is a chronic, diffuse abnormality of the meibomian glands, commonly characterized by terminal duct obstruction and/or qualitative/quantitative changes in the glandular secretion’ and is the leading cause of DED.[9] The structure of meibomian glands is evaluated using meibography, and lipid layer thickness (LLT) and meibum quality or expressibility represent its function.[36-38] The following sections mainly discuss the clinical utility of these diagnostic techniques.

Meibum quality and expressibility

The functional status of meibomian glands is assessed by testing the expressibility of meibum and the quality of expressed meibum [Fig. 4]. Meibum quality is graded as 0 (clear fluid), 1 (cloudy fluid), 2 (cloudy particulate), or 3 (toothpaste-like). The expressibility of meibum from the central eight glands of the upper and lower eyelids is graded as 0 (all glands expressible), 1 (3 to 4 glands expressible), 2 (1 to 2 glands expressible), and 3 (no glands expressible). The amount of pressure applied for expression varies across studies. Some report firm digital pressure, and some have reported standardized digital pressure.[9] There are also meibomian gland expressors and evaluators available in the market; however, authors have observed that patients report discomfort when these devices are touched onto the tarsal surface of the eyelid. The meibum quality and expressibility have shown a good correlation with gland loss and low LLT values.[37-39] However, ocular surface staining and TBUT showed no significant correlation with meibum quality in obstructive MGD patients.[40] The meibum grade should be evaluated in both upper and lower eyelids clinically as the mean grade of meibum quality can differ significantly between the two.[38] In contact lens users, meibum quality did not correlate with gland dropout in the lower eyelids but correlated in the upper eyelids.[41] Also, the lid margin is examined for vascular engorgement, irregularity, displacement of Marx line, or plugging of orifices [Fig. 4].

Figure 4.

(a) Lid margin of a normal adult showing clear meibum quality when expressed from the upper eyelid. (b) Thick, whitish, and cloudy meibum expressed from a patient with MGD. (c) Increased lid margin vascular congestion in a patient with MGD with inflammation around meibomian gland openings

Meibography

Non-contact infrared meibography was introduced in 2008, and it provides a two-dimensional structure of meibomian glands [Fig. 5]. Normal meibum shows autofluorescence and is seen as light areas on infrared meibography. Numerous studies have described the structural changes observed on meibography in DED patients.[36-45] Different meibographic morphologies are a deviation from normal vertically running glands. They have been divided into variations in length (short, dropout), width (thick, thin, overlapping), or gland shape (tortuous, hooked, distorted).[42] Gland loss has been the most commonly studied entity, and different grading methods are available for describing gland loss areas. Meibography grading systems are based on the visible area of the tarsal plate without glands (meiboscore <33%, 33–66%, >66%) or the number of partial glands or dropout glands (0, <3 or more than 3) and image J-based analysis of dropout area (<25%, 26–50%, 51–75% or >75%).[36] The suitable objective measures for interpreting meibography like artificial intelligence-based are still at the research stage. Gland dropout or loss areas do not have glandular tissue and appear dark on meibography. Gland dropout and short glands significantly correlate with severe DED symptomatology.[39-42,44,45] In very few studies, gland tortuosity has also shown predominance in patients with early MGD and allergic conjunctivitis.[43] The structure–function correlation of meibographic morphologies has shown no significant difference between the morphologies histologically, except for severe short glands.[7] Severely short glands (<1/3^rd length) demonstrate loss of meibocyte differentiation and cellular proliferation. In routine clinical practice, a simple assessment of gland loss areas and the presence of short glands indicate meibomian gland involvement in DED. It might not correlate well with meibum expressibility or lipid layer thickness.[44] Hence, meibography alone cannot be used. When 2 of the 3 scores (symptom score, lid margin abnormality, and meiboscore) were abnormal, the sensitivity was 84.9% and the specificity was 96.7% in diagnosing obstructive MGD.[40] Another way of imaging MGs is OCT meibography, a customized system under research. It provides a 3D image of meibomian gland acini and ducts with a better resolution.[46] OCT meibography showed a decrease in the meibomian gland length and width compared to the control; however, the number and diameter of meibomian gland orifices were similar between MGD and control eyes. In vivo confocal microscopy (IVCM) can also provide acinar details similar to histological methods, but evaluation needs to be standardized with image analysis software.[47] The described parameters on IVCM are meibomian gland acinar unit density, the longest and shortest acinar diameters, and inflammatory cell density with cut-off values of 70 glands/mm², 65 μm, 25 μm, and 300 cells/mm², respectively. A standardized meibography reporting system should be devised for ensuring uniform evaluation across studies.

Figure 5.

A1 & A2, Infrared meibography of upper and lower eyelids of a normal adult showing vertically running meibomian glands reaching up to convex tarsal border. In DED patient, there is gland loss involving 1/3^rd to 2/3^rd of tarsal area (B1 & B2)

Lipid layer thickness

The specular reflection at the junction of lipid–aqueous layer produces interference patterns, which are analyzed quantitatively or semi-quantitatively, indicating lipid layer status.[48-53] The various non-invasive diagnostic devices for assessing lipid layers are Tearscope (Keeler, Windsor, UK), DR-1 α (Kowa, Nagoya, Japan), IDRA ocular surface analyzer (SBM Sistemi, Coburn Technologies, Inc, Italy), and Oculus keratograph (K5M; Oculus®, Wetzlar, Germany) or LipiView interferometer (TearScience Inc, Morrisville, NC, USA). Tearscope is attached to a slit lamp, and different patterns of lipid layer appearance are given a range of thicknesses. LipiView and IDRA-OSA give better LLT repeatability than Tearscope. The DR-1α interferometer measures the light reflection from the lipid–aqueous interface. It shows colored patterns of the tear film with blink and provides tear film kinetics, its spread, and the wettability of the ocular surface.[53] The test results are graded subjectively. Similarly, a keratograph gives a lipid layer video with different colored patterns of tear spread, and no standardized system exists so far for interpreting them. LipiView and IDRA-OSA give a numerical value to LLT. One interferometric color unit is equal to 1 nm. Values less than 60 nmm are considered pathological. LipiView provides minimum, maximum, and average LLT along with a number of complete and partial blinks [Fig. 6]. The LLT values using LipiView II interferometer range from 33 to 100 nm in normal individuals.[49] LLT values are affected by age in normals and increase with increasing age. The coefficient of variability for inter-observer repeatability is 13 nm and the intra-observer repeatability 16 nm in healthy individuals.[52] The sensitivity and specificity of LipiView interferometer are 65.8% and 63.4%, respectively, if a cut-off value of 75 nm is used for MGD diagnosis.[37] The reported LLT in MGD patients is 79 nm (20–100) and is affected by the history of cataract surgery, refractive surgery, older age, sex, Schirmer values, and lid margin inflammation.[49] Hence, LLT should be carefully interpreted, taking into account other factors. LLT values obtained using LipiView correlate well with MG loss and MGD grade.[38,50] No correlation exists between NIBUT and LLT values or TMH and LLT values measured with any device.[48]

Figure 6.

Lipid layer thickness output from a LipiView interferometer giving minimum, average, and maximum thicknesses along with blink rate

LLT values show good repeatability measured with either Lipiview or IDRA-OSA in both healthy and DED patients.[54] The CoV is more in DED patients (IDRA, 26.96, LipiView, 37.57) compared to the control (IDRA,19.66, LipiView, 36.85).[54] Between the machines, no significant difference in LLT was observed between LipiView II and IDRA in 47 non-Sjogren’s DED subjects.[52] The reported average LLT in this study was 77 nm (SD, 24.2) using LipiView and 75.3 nm (SD, 13.01) using IDRA®, which is more than the cut-off for DED (<75 nm). These diagnostic devices require a trained operator, and values can significantly differ between two observers; hence, the same person should be performing the test in clinics.

Tear break-up time

Tear film break-up time (TBUT) measures tear stability, essential for maintaining optical clarity. It captures the time taken for the precorneal tear film to break after one blink.[55-59] TBUT is affected by tear film viscosity and surface tension, which are affected by tear film composition. Any corneal epithelial irregularities, reduced lipids in the tear film, the altered thickness of the aqueous layer, and diminished mucus layer attachment to corneal microplicae would affect TBUT. It can be measured using fluorescein dye or non-invasively using the tear film interferometry.[55-62] The tear film has inherent physiological variation; hence, the values obtained with either instrument show not-so-good repeatability and reproducibility.

Fluorescein-based TBUT

In a dimly lit room, a drop of fluorescein dye from a strip or ten microliters of dye from unims is instilled into the conjunctival sac without touching the ocular surface. The excess saline should be shaken off the strip, and a smaller segment should be moistened. The subject is instructed to place the chin over the slit lamp and blink 2–3 times. Then, the whole precorneal area is observed for a first random dry spot (black area) while the patient is instructed not to blink. The accurate calculation of the time can be performed using a stopwatch. The patient should be counseled to keep his eyes open rather than holding the eyelid forcefully. Touching the eyelid can induce reflex tearing. The tear film is the thinnest in the lower temporal area, and it usually shows up first. Fluorescein concentration and observer-based differences make this test variable. The reported mean FBUT values in normal subjects were 7.6 ± 10.4 s.[55] If the TBUT is less than 10 s, either invasive or non-invasive, one should repeat the test to label it pathological. Look out for the area with dry spots if it is the same with repeated evaluation, which might indicate epithelial irregularity in that area. With the advent of non-invasive tests, the accuracy of TBUT is better when tested two or more times. The palpebral aperture can impact TBUT values as opening eyes wide open might break eyelid margin touch with tear lake. Any drops instilled before the test, like paracaine, lubricants, and Schirmer strip use, can also affect the result.

Non-invasive TBUT

Non-invasive TBUT (NIBUT) measures changes in reflected videokeratographic mires using topographic systems. The Placido disc projects the mires onto the cornea, and any changes in the mires’ regularity are taken as the onset of tear break-up. Oculus keratograph gives the first break-up time as NIBUT-1 and the average break-up time for all break-up points as NIBUT-average [Fig. 7].[59] NIBUT values are lower than FBUT as fluorescein dye impacts the TBUT values. The mean difference between invasive and non-invasive techniques is up to 3.7 s, which tends to go lower for lesser values of NIBUT.[55] The average NIBUT (Oculus®) reported in healthy subjects is between 10 to 11 s with an SD of 3.9 to 5.2 seconds.[48,59,62] In a normative data study of 236 individuals of South Indian origin, 33% had NIBUT values between 7 to 10 s and 7% had less than 7 s.[61] Age and sex do not affect TBUT values unlike meibography.[61] NIBUT shows a good correlation with DED symptomatology and correlates significantly with TMH in DED patients.[60,61] An average of three NIBUT measurements is considered more accurate. A cut-off value of fewer than 10 s has been taken as a diagnostic level for EDE.[8] A 10 s cut-off gives 82% to 84% sensitivity and 76% to 94% specificity.[8] Of all non-invasive tear film parameters, an NIBUT value of less than 7.7 s had 89% sensitivity and 69% specificity in diagnosing mild to moderate DED, whereas LLT, TMH, or gland loss on meibography had a sensitivity between 50 to 60%.[60] NIBUT using K5M and IDRA-OSA shows good repeatability and reproducibility in the healthy subjects but poor repeatability in DED patients.[62] Between the two machines (K5M and IDRA-OSA), the keratograph shows higher NIBUT values than IDRA-OSA, and no agreement exists between the two devices.[48] The CoV values are higher in DED patients compared to normal individuals with both K5M (28 vs. 47.7%) and IDRA-OSA (25 vs. 48%).

Figure 7.

Non-invasive tear break-up time captured with Oculus keratograph in a patient with meibomian gland dysfunction

Recommendations in Clinical Use

There is no single diagnostic test of lacrimal or meibomian gland damage in DED. A combination of tests are performed for diagnosing the DED sub-type. One tear film test usually interferes with other tests as maintaining eye opening for NIBUT assessment, eyelid eversion for meibography, and reflex tearing because staring into the light source is expected to alter tear dynamics. The recommended sequence of tests for diagnosing DED is a dry eye questionnaire followed by non-invasive tests such as TMH, LLT, NIBUT (FBUT if non-invasive test unavailable but should be performed after osmolarity), and tear osmolarity, followed by invasive tests such as meibography, ocular surface staining, and meibomian gland expressibility and quality.[8] Schirmer test should be performed after non-invasive tear film diagnostic testing and after fluorescein was washed off. One should use a standardized way of measuring the tear film parameters in clinics, and preferably, the same person should perform the diagnostic tests at every visit to reduce inter-examiner variability. Tear film parameters display heteroscedasticity, which means that more significant variations are seen at greater values.

Future Directions

Non-invasive tear film and ocular surface imaging allows objective measurements of the tear film in DED patients. The detailed tear film analysis has improved the DED diagnosis and management as well as reporting the clinical trials’ outcomes. The cut-off values of different devices and parameters still need consensus and more repeatability studies. Although much research focused on non-invasive tear film assessment, the accuracy in differentiating the severity of DED and defining management options is still uncertain. Further studies exploring the impact of these investigations on patient-reported outcomes would be of clinical relevance. The cost benefit ratio of these diagnostic devices into our clinical practice could be justified if DED diagnosis and management are based on guidelines established using different modalities.

Financial support and sponsorship

The authors have no proprietary or commercial interest in any materials discussed in this manuscript. FP receives royalties from Elsevier for the 24^th and 25^th Ed. of the anatomy atlas “Sobotta” and for the “Sobotta Textbook of Anatomy”. FP was also supported by Deutsche Forschungsgemeinschaft (DFG) grant PA738/15-1. The funding organization had no role in the design or conduct of this research. Dr Swati Singh is supported by Alexander-von-Humboldt-Foundation, Germany.

Conflicts of interest

There are no conflicts of interest.