Case Study 18.1 — Approved There, Not Approved Here

Thymosin alpha-1 and what a split regulatory decision actually tells you

The situation

A compound has been available as a prescription medicine in a number of countries for a long time. It is a 28-amino-acid peptide, thymosin alpha-1, derived from prothymosin alpha, with immunomodulatory activity. In the jurisdictions where it is approved, it carries defined indications — hepatitis B is the most prominent, and in some places it is approved as an immune adjuvant.

It is not FDA-approved in the United States.

The evidence base those regulators looked at is, in broad terms, the same body of work. It includes randomized human trials — more human trial evidence than any other compound in this chapter, by a wide margin. It also includes trials that differ considerably from one another in design, population, endpoint, and quality, and reported effect sizes that are contested.

So: same molecule, largely the same evidence, different institutional conclusions.

This is the entire case study. There is no scandal to uncover, no suppressed document, no corrupt official. The interesting thing is exactly what it appears to be, and most commentary manages to miss it by insisting that somebody must have made a mistake.

The two stories people tell

Story one — the promotional reading. It's approved in a whole list of countries. Regulators there have looked at the evidence and cleared it. The FDA is slow, or captured by the companies whose products this would compete with, or excessively conservative in a way that costs patients. American patients are being denied something the rest of the world has.

Story two — the dismissive reading. The FDA has the highest standards in the world and it has not approved this. Other agencies have lower thresholds, less capacity, or more political exposure. If it worked, the FDA would have approved it. The approvals elsewhere are noise.

Both stories are internally coherent. Both are told by serious people. And both rest on the same unexamined premise: that one regulator is right and the other has erred.

Set that premise down and pick up a different one. Suppose both decisions were reasonable, made in good faith, by competent people applying defensible decision rules to the same material. What would have to be true about the evidence for that to be possible?

What the split implies

Regulators are institutions applying decision rules under uncertainty. They differ in the thresholds they set, the endpoints they accept as adequate, the statutory mandates they operate under, the comparator therapies already available in their markets, and the disease burden their populations carry. A jurisdiction where a disease is common, severe, and inadequately treated faces a different benefit-risk calculation than one where the same disease is rarer and better served by existing options. That is not corruption. It is what a benefit-risk calculation is.

Now run the three possibilities.

If a compound's effect were large, consistent, and measured on unambiguous endpoints, you would expect regulators to converge. Different thresholds would not matter much, because the evidence would clear all of them. That is what approval of insulin, or of the GLP-1 agonists for their labeled indications, looks like.

If a compound had no credible human evidence at all, you would expect no approvals anywhere — usually because nobody has even assembled a dossier. That is the situation for most of the compounds in Part III of this book.

The remaining case is a split. Some clear it, some do not. And the natural reading of a split — the one that does not require anybody to be incompetent — is that the evidence is genuinely strong enough to satisfy some reasonable decision rules and genuinely weak enough to fail others. The disagreement is not noise in the institutions. It is signal about the evidence.

Why this is uncomfortable

Readers want verdicts. "The evidence is genuinely contestable" reads as a dodge, a diplomatic non-answer, the kind of thing written by someone who does not want to offend either side.

It is not that. It is a positive claim with content, and the content is: there is real human evidence here, and it does not settle the question. That is a different statement from "we don't know," and very different from "there is nothing to look at." It says a specific thing about where a specific compound sits.

This book's ⚠️ is that statement. Most Part III compounds are ❌ because there is nothing to weigh. Thymosin alpha-1 is ⚠️ because there is something to weigh and the scale has not tipped. It is the book's clearest example of the genuinely contestable middle, and if the rating system cannot represent that position honestly, the system is not a rating system — it is a device for sorting compounds into the pile their advocates or detractors already wanted them in.

What this does and does not change for a person

The temptation, having established that the evidence is contestable, is to conclude that the practical question is therefore a toss-up. It is not, and the reason is that regulatory status carries practical consequences independent of the evidence question.

A compound that is not approved in a jurisdiction is, in that jurisdiction, not an approved medicine — which has implications for how it could be obtained, what quality assurance sits behind what arrives, and what recourse exists if something goes wrong. Chapter 35 covers that terrain, and the quality question is not a footnote. An approval also attaches to an indication: an approval for chronic hepatitis B is an approval for chronic hepatitis B, in patients with chronic hepatitis B, and it is not an approval for general immune support, for recovery from training, or for prevention of infection in someone healthy.

And an approval does not report an effect size. It reports that a regulator judged benefit to outweigh risk for a defined use. Whether the benefit was substantial or marginal is a question for the trials, not for the approval letter — and the trials, in this case, are exactly where the contest lives.

Discussion questions

1. Both the promotional and the dismissive readings share a hidden premise. State it precisely. Then describe what evidence would be needed to justify that premise in a particular case — that is, what would actually establish that one regulator had erred rather than merely reached a different defensible conclusion?

2. The case study argues that convergent approval, universal non-approval, and split approval each imply something different about the underlying evidence. Construct a counterexample: a plausible scenario in which a split occurs for reasons that have nothing to do with the strength of the evidence. How would an outside observer distinguish your scenario from the one described here?

3. Disease burden affects benefit-risk calculations, so a compound can be reasonably approved where a disease is common and reasonably declined where it is rare. Is that a defect in the idea of drug approval, or a correct feature of it? Argue whichever side you find less comfortable.

4. This book defends ⚠️ as a positive claim rather than a hedge. Design a test: what would you look for in a source's body of ratings to determine whether its middle category carries real content or is functioning as a place to put anything controversial?

5. A patient in a country where thymosin alpha-1 is not approved asks whether the approvals elsewhere mean they should try to obtain it. Identify every distinct question buried in that request — there are at least four — and say which ones this case study can answer, which belong to Chapter 35, and which belong to a clinician who knows the patient's history.

6. Compare thymosin alpha-1's ⚠️ with TB-500's ❌. Both compounds are promoted in overlapping markets, sometimes on the same page. Write the paragraph you would give a reader who wants to know why these two are being treated so differently — without using the word "evidence" more than twice, and without implying that the ⚠️ compound is therefore a good idea for them.