When IPL Doesn't Work: What Predicts a Poor Response

Intense pulsed light (IPL) is one of the more widely used in-clinic treatments for meibomian gland dysfunction (MGD), where the oil glands in the eyelids become blocked or inflamed. But not everyone feels better after it. A 2026 study from Seoul, South Korea, looked at people who had already had IPL without enough relief, gave them more IPL, and asked what set apart those who improved from those who didn't.

What the study did

This was a retrospective study, meaning the researchers reviewed existing medical records rather than setting up a trial in advance. It took place at a single hospital, Asan Medical Center in Seoul, and included 66 adults (one eye each) with moderate-to-severe MGD. All were still symptomatic after four IPL sessions using a 590 nm filter at other clinics, and had not improved with steroid and cyclosporine drops, warm compresses and lid scrubs.

They then received four further IPL sessions, two weeks apart, using an M22 IPL device (Lumenis) fitted with a "vascular" filter (which passes light at 530–650 nm and 900–1200 nm). Each session was followed by gently expressing the glands of all four eyelids. Throughout, everyone also used loteprednol steroid drops four times a day, cyclosporine drops once a day, a lubricating gel and drops, warm compresses and lid scrubs. There was no comparison group that did not receive the extra IPL.

How "response" was decided

One month after the final session, patients rated their symptoms compared with before treatment as moderately improved, mildly improved, unchanged or worse. Those reporting moderate or mild improvement were called responders (32 people); those reporting no change or worsening were called refractory (34 people). This was a single, study-specific rating that the authors describe as non-validated, and it was not measured at the start. Even so, it gives a useful sense of expectations: roughly half of these previously treated patients did not feel better after four more sessions.

What set the two groups apart

Before the extra treatment, the two groups did not differ significantly on any measure except one: staining of the eye surface. Staining uses a dye that highlights damaged surface cells, and a higher score means more damage. On the Oxford staining scale, responders scored 2.5 ± 1.1 and the refractory group 3.3 ± 1.4 (P = 0.020). A P-value of 0.020 means a difference this large would be fairly unlikely if the groups truly did not differ.

One month after treatment, responders had less staining on both the Oxford scale (1.1 ± 0.3 vs 2.1 ± 1.1; P < 0.001) and the SICCA scale (2.1 ± 1.9 vs 4.9 ± 2.4; P < 0.001). They also showed greater improvement in staining, oil quality and consistency, and gland grade. The size of those changes was in tables we could not access. Tear break-up time was similar between groups (3.3 ± 1.1 vs 3.0 ± 1.2 seconds; P = 0.199), and so was MMP-9, a marker of surface inflammation (1.8 ± 1.4 vs 2.2 ± 1.0; P = 0.202). That is worth noting because the authors' discussion suggests poor responders had high MMP-9, yet the reported difference was not statistically significant.

What the studies can't tell us

• Whether the extra IPL caused any improvement. With no control group, and everyone also using steroid and cyclosporine drops and lid care, the effect of IPL can't be separated from the other treatments or simply time.
• Whether "response" means what it sounds like. Groups were defined by one self-rating at one month. The authors themselves caution that "refractory" patients were not necessarily treatment failures, as some had objective improvement without feeling better.
• How reliable the predictor is. The staining finding comes from a simple comparison of 66 people, with no adjustment for other factors and no correction for multiple comparisons. The sample size was not planned in advance.
• Longer-term outcomes and safety. Follow-up was one month, and side-effects were not reported for this group.
• The full picture. Details of the earlier IPL at other clinics were unknown, gland loss (atrophy) was not imaged, and funding and conflict-of-interest statements were not in the text available to us.

What this means for you

The most useful message is a practical one. In this study, people whose eye surface was more damaged before treatment were less likely to report feeling better afterwards. While this can't prove cause and effect, it fits with a principle we see often: if the surface is significantly inflamed or damaged, that may need attention alongside work on the oil glands. It is another reason to treat the cause, not the symptom, and to start with a thorough assessment rather than a single device.

A specialist dry eye workup looks at the glands, tear film, eye surface, lids and contributing factors, so a plan can be built around what is actually going on. IPL therapy can be a helpful part of that plan for some people, but it is one tool among several, and results vary. At Rose Optometry in Hamilton, a DESG member practice, the optometry team, including Jagrut Lallu and Jacqueline Rowe, offers IPL as part of its Dry Eye Clinic, starting with a paid Dry Eye Evaluation. You can also find other DESG member practices around New Zealand.

References

• Jeon YY, Park N, Lee YE, et al. Characteristics of patients with refractory meibomian gland dysfunction unresponsive to intense pulsed light treatment using a vascular filter. Diagnostics (Basel). 2026;16(16):2532. PMID: 42650935. https://doi.org/10.3390/diagnostics16162532