Meibography

The meibomian glands are long glands located inside the upper and lower eyelids. Because they are hidden within the eyelid tissue, we cannot see their full structure simply by examining the lid margin.

Meibography uses imaging — most commonly infrared light — to make the glands visible through the eyelid. The result is essentially a two-dimensional image of the gland structures that the imaging system can detect.

A clinician can look at how many glands are visible, their length, their shape, their distribution, and areas where normal gland structures are not visible. This gives us valuable information about gland morphology.

But it is important to remember: meibography is an image, not a direct test of gland secretion.

One of the most common terms used in meibography is gland dropout. You may also hear gland loss.

TFOS DEWS III describes meibography as imaging the two-dimensional silhouettes of the meibomian glands and defines dropout or loss as areas of partially or totally non-visible meibomian tissue.

That wording is important. Meibography tells us that normal gland tissue is not visible on that image.

Clinicians or imaging software can estimate how much of the evaluated gland area appears non-visible and may report this as a percentage or as a meiboscore. More non-visible area generally indicates greater apparent structural gland loss. But the image still needs interpretation.

This is where we need to be particularly careful.

An area that appears as dropout on meibography should not automatically be translated into “Every gland in this area is completely absent and incapable of producing secretion.”

The imaging system is showing us which gland structures it can visualize. It is not directly measuring secretory activity. And the appearance of the image can be influenced by technical factors. For example: image quality matters, how completely the eyelid is everted matters, which part of the eyelid is captured matters, the device matters, and the image-analysis method matters.

So “non-visible on this image” and “biologically proven to be completely absent and non-functional” are not identical statements.

Meibography itself cannot answer this question.

If secretion can be obtained from a gland opening, that gives us functional information. If a gland structure is not clearly visible on meibography, that gives us imaging information. If those findings appear not to match, we should not automatically assume that one of them must be wrong. Instead, we need to ask: was the gland adequately visualized? Was the entire relevant part of the eyelid captured? Was image quality good? Which gland opening produced secretion? Can that opening confidently be matched to the structure seen — or not seen — on the image?

Current evidence does not justify saying that every apparently dropped-out gland is actually intact and functioning. But the reverse conclusion also goes too far: an image alone does not directly prove the functional status of every gland in an apparently dark area.

That is why structural and functional findings should be interpreted together.

This is perhaps the most useful way to understand the relationship between these two tests.

Meibography asks: what gland structures can we see, and what do they look like? Meibomian gland expression asks: what secretion can we obtain from the glands under the conditions of this examination?

meibography → structure→expression → function

Neither provides the complete answer. And when the two appear not to match, that discrepancy can itself be useful information. It should prompt closer interpretation rather than an immediate verdict.

Meibomian Gland Expression & Meibum Quality

Potentially, quite a lot. This is not just a theoretical limitation.

In one study, researchers performed lower-eyelid meibography on the same 20 people during the same visit using two widely used infrared imaging systems: LipiView II and Keratograph 5M. When the images were analyzed for percentage gland dropout, the average result was 31.5% dropout with LipiView II compared with 43.4% dropout with Keratograph 5M.

These were the same people and the same lower eyelids — yet the average estimate differed by almost 12 percentage points. For individual measurements, the disagreement could be much larger. The statistical limits of agreement between the two systems extended from approximately 43 percentage points lower to 19 percentage points higher. Even the categorical meibography score could differ by as much as two grades between the devices.

The researchers attributed these differences partly to image quality and to how much of the eyelid was successfully everted and captured by each system. Their conclusion was important: the two devices should not be used interchangeably when measuring meibomian gland dropout.

This issue has not disappeared with newer technology. A much larger study involving 227 participants, published in 2025, again compared LipiView II with Keratograph 5M. Both systems had useful diagnostic performance. But their percentage-dropout measurements still showed wide limits of agreement — roughly 25–30 percentage points in either direction, depending on the eyelid being measured. The researchers again concluded that measurements from the two systems were not directly interchangeable.

Other comparisons between infrared meibography systems have also reported meaningful differences.

This tells us something important: a meibography percentage can look extremely precise on a report, but the number is partly dependent on how the image was obtained and analyzed.

Imagine that after your examination you are told: “You have 40% gland dropout.”

It is very easy to hear this as “Exactly 40% of my meibomian glands are physically gone.” Or even “Forty percent of my glands are permanently dead.”

But that is not what the measurement establishes with that level of precision. The percentage is an estimate derived from the gland structures visible on that particular image, the area of eyelid successfully captured, the imaging system used, and the method used to define and calculate the apparent dropout area.

This does not mean that meibography is unreliable or unimportant. It means that the number needs to be interpreted as an imaging measurement, not as an exact count of permanently lost functioning glands.

This distinction matters because patients may carry these numbers with them for years.

A result such as “40% gland dropout” sounds final. But a more accurate way of understanding it is: “On this image, using this device and this analysis method, approximately 40% of the evaluated gland area was classified as gland dropout.”

That is still clinically useful information. But it is a much more accurate description of what was actually measured. And it leaves room for the other information that matters: what the glands actually look like, what can be expressed from them, what the meibum quality is, what is happening at the lid margin, what is happening in the tear film, and whether repeated examinations tell a consistent story.

Meibomian glands normally appear as elongated structures running through the eyelid toward the lid margin. But normal anatomy is not perfectly uniform. Upper and lower eyelids are different. Upper-eyelid glands are generally longer and more numerous. Individual glands can also vary somewhat in shape.

So healthy meibography does not mean every gland must look like a perfectly straight, identical line. The overall pattern matters more than whether one individual gland looks unusual.

Some glands may appear shorter than expected. This can be described as gland shortening or truncation.

Gland length appears to be clinically meaningful. Current TFOS DEWS III evidence suggests that gland length is one of the morphological characteristics that may relate particularly closely to gland expressibility.

But even here: structure is not the same thing as function. A short gland is a structural finding. Expression is needed if we want information about what secretion can actually be obtained from it.

Some glands appear unusually curved, twisted or wavy. This is called tortuosity. Tortuosity has been studied in relation to MGD, tear-film stability and dry-eye symptoms.

But its meaning is not straightforward. Current evidence does not support using gland tortuosity by itself as a reliable diagnosis of dry eye or as a complete measure of gland function.

So a curved gland is a morphological finding — not a diagnosis.

Meibography can reveal other morphological changes too. Glands may appear dilated, thinned, distorted, shortened or irregular. These findings may contribute useful information about the gland system. But no single morphological feature should be interpreted as a complete diagnosis or prognosis. Current TFOS DEWS III evidence indicates that individual morphological characteristics considered alone are not good predictors of dry eye disease.

Yes. This is extremely important.

Meibomian gland dropout can also be found in people without dry-eye symptoms. Gland morphology changes with age and varies between individuals.

So seeing some dropout on an image does not automatically mean “I have severe MGD.” And it certainly does not automatically mean “I will develop severe dry eye.” The image needs clinical context.

Not necessarily. Structural gland abnormalities and symptoms do not have a simple one-to-one relationship.

Someone can have substantial visible structural changes and relatively few symptoms. Another person can have severe symptoms with much less dramatic meibography.

This is a recurring theme throughout dry-eye diagnosis: what we can see, what we can measure and what a patient feels are related — but they are not interchangeable.

No. A clearly visible, structurally preserved gland is reassuring from an anatomical point of view. But the image cannot prove that the gland produces normal meibum, delivers it through the duct normally, has a functioning opening, or contributes normally to the tear film. Those are functional questions. A gland can therefore look relatively preserved on imaging while expression is abnormal.

Again, not necessarily. Structural abnormalities may be associated with impaired function, and some morphological measurements correlate with expressibility. But morphology does not directly measure secretion. That is why it is useful to ask both: what does the gland look like, and what can it actually deliver?

Ideally, yes. Upper and lower eyelids have different gland anatomy, and abnormalities may not be distributed equally between them. TFOS DEWS III specifically emphasizes imaging both upper and lower eyelids when assessing gland dropout. So an image of only one lower eyelid does not necessarily represent the entire meibomian gland system.

Meibography is an important part of MGD assessment. But it should not be used alone. A clinician may also need to assess meibum quality, gland expressibility, gland openings, lid-margin abnormalities and other functional findings. Meibography contributes the structural information. It does not replace the rest of the assessment.

No. A meibography image does not diagnose dry eye disease. A person can have structural meibomian gland abnormalities without meeting the diagnostic criteria for dry eye. And someone can have dry eye even when meibomian gland structure is relatively well preserved.

Meibography is therefore useful for helping identify meibomian gland dysfunction as a possible contributor to the larger dry-eye picture. So gland dropout is not a dry-eye diagnosis.

Repeated meibography can potentially help monitor structural changes over time. But comparisons need to be made carefully. Ideally, follow-up images should use the same device, similar eyelid eversion and positioning, similar image quality, and the same analysis method. Otherwise, an apparent change may partly reflect the imaging process rather than a true biological change.

For example, imagine one examination reports 25% dropout and a later examination reports 40%. Before concluding “I lost another 15% of my glands,” ask: was the same machine used? Was the same area of the eyelid captured? Was image quality comparable? Was dropout calculated in the same way? And then look at the actual images and functional findings. Small differences between two measurements should not automatically be interpreted as disease progression.

Not precisely. A meibography image gives us useful information about gland morphology now. But it cannot reliably tell an individual patient how severe their symptoms will become, whether their condition will progress, how quickly progression might occur, whether a particular treatment will work, or what their eyes will feel like years from now.

This matters because meibography images can look frightening. A structural image is valuable clinical information. It is not a forecast of your future.

Meibography can help show:

• •which meibomian gland structures are visible
• •their length and shape
• •areas classified as gland dropout or loss
• •other morphological abnormalities
• •how those structural findings are distributed across the eyelids

This makes it an important tool for understanding meibomian gland morphology.

By itself, it cannot tell you:

• •how much meibum a gland is producing
• •whether meibum can reach the lid margin
• •what the secretion looks like
• •whether every apparently non-visible gland is completely incapable of secretion
• •how severe your symptoms should be
• •whether you have dry eye disease
• •exactly what will happen to your glands in the future

Those questions require other information.