Thrombosis and Extracorporeal Photopheresis: What the Risk Actually Looks Like

Extracorporeal photopheresis (ECP) has one of the best safety reputations in procedural medicine. It’s been used for decades. Hundreds of thousands of treatments. Indications ranging from cutaneous T-cell lymphoma to chronic graft-versus-host disease. And yet, every so often, the same question resurfaces:

Does ECP increase the risk of thrombosis?

The short answer is: there is a signal, but it’s small, context-dependent, and often misunderstood. The longer answer is more interesting—and more useful.

Where the concern comes from

In 2018, the FDA issued a letter to healthcare providers warning of reported cases of venous thromboembolism (VTE), including pulmonary embolism, in patients undergoing ECP with the THERAKOS CELLEX system.

That sentence alone has done a lot of quiet work over the years.

What often gets lost is why the FDA issued the letter and what it actually said.

The warning was based on post-marketing reports, not on prospective trials or large cohort studies. The FDA described seven pulmonary emboli and two deep vein thromboses, all occurring in patients treated for chronic GVHD. Two of the pulmonary emboli were fatal. The mean time to event was about 1.7 days, leading to the phrasing that events occurred “during or shortly after” treatment sessions.

Importantly, the FDA did not conclude that ECP causes thrombosis. The language was careful: ECP may increase risk, based on timing and clustering in a vulnerable population.

That distinction matters.

What the published literature shows (and doesn’t)

If you go looking for thrombosis in the ECP literature, you’ll find… very little.

Across more than 30 years of published experience:

• Thrombotic events are rare

• Most reported cases are catheter-associated, not systemic

• Large case series and reviews consistently emphasize ECP’s excellent safety profile

• Coagulation parameters remain stable during treatment, even with long-term therapy

• Laboratory studies show platelet activation after UVA/8-MOP exposure—but without aggregation or downstream thrombotic effects

  
  

In pediatric cohorts, multicenter studies, and long-term follow-up reports, thrombosis appears as an isolated complication, not a recurring pattern.

That doesn’t mean the FDA signal was wrong. It means the signal exists in a space the literature hasn’t fully interrogated.

The missing denominator problem

One of the hardest things about post-marketing safety signals is that they arrive without context.

We don’t know:

• How many total ECP treatments occurred during the reporting window

• Whether events clustered around central venous access

• How immobility, inflammation, infection, or baseline hypercoagulability contributed

• Whether similar patients not receiving ECP had comparable short-term VTE rates

  
  

And chronic GVHD patients—who made up all reported cases—already carry a high baseline risk of thrombosis.

When a population is fragile enough, even a neutral intervention can appear suspicious if you look only at timing.

So where does that leave us?

A reasonable, evidence-based position looks something like this:

• ECP is not a high-thrombosis procedure

• There is a small regulatory safety signal, concentrated in a very high-risk population

• Timing alone does not establish causality

• Access-related thrombosis likely explains a meaningful fraction of reported events

• Clinicians should remain alert—but not alarmist

  
  

This is not a story of a dangerous therapy being uncovered. It’s a story of how safety signals emerge, how they should be interpreted, and how nuance gets flattened over time.

Why this matters

ECP is often used when options are limited. Overstating risk can quietly narrow access to a therapy that is otherwise well-tolerated and effective.

At the same time, ignoring regulatory signals entirely isn’t good medicine either.

The work, as always, is in the middle: understanding who might be at risk, when vigilance matters most, and how to contextualize rare events without letting fear do the thinking.

Bottom line:

If thrombosis were a common or intrinsic complication of ECP, we would know by now. What we have instead is a small, signal-level warning that deserves clarity—not amplification.

And clarity is something we can still build.