What Your Doctor Cannot See

Compounded semaglutide and tirzepatide appear in 8.2% of primary care records for people taking these drugs. Surveys of patients put the real figure near 23%.^[1] If both are right, roughly two-thirds of compounded use is invisible to the medical record — and everything that depends on the record fails quietly along with it.

Why the gap exists

Researchers examined a nationwide database of primary care records covering 153,044 patients with documented semaglutide or tirzepatide use between 2021 and 2024. Brand-name prescriptions were straightforward to find — they sit in structured prescribing data.

Compounded use had to be extracted from clinical notes. There is no prescription record in the system, because the drug did not come through it. It exists in the chart only if the patient mentioned it and a clinician typed it out.

A medication bought outside the system leaves no trace inside it unless someone writes one.

What stops working

• Drug interactions. These medicines slow gastric emptying, which affects how other oral drugs are absorbed — covered in your other medications. A checker cannot flag a drug it does not know about.
• Surgery and endoscopy. Retained stomach contents despite fasting are roughly six times more likely on these drugs, which changes how a procedure is prepared for — see before endoscopy and surgery.
• Side-effect attribution. A symptom is far less likely to be connected to a medication nobody has recorded.
• Safety surveillance. If two-thirds of compounded use is invisible, so is two-thirds of whatever it causes.

None of that requires anyone to have done anything wrong. It follows from a supply route that runs around the system rather than through it.

Who is actually using it

The non-diabetic part fits that reading. Someone with type 2 diabetes has a covered indication and a prescriber writing brand-name product; someone seeking weight treatment without a covered indication is paying either way, and a cheaper cash price is a smaller obstacle than an insurance denial.

One caution on how far to take this. Socioeconomic deprivation here is measured by area, not by individual income — a neighborhood-level proxy that describes where someone lives rather than what they earn. It is a real signal and a blunt one.

The study also found compounded-only users stayed on treatment longer, averaging 10.0 months against 7.8 for brand-only users. We are not going to explain that confidently. It could be uninterrupted cash-pay supply against insurance that lapses, or a difference in who chooses which route, or something the data cannot see.

What follows from it

The actionable part is short. If you are taking a compounded GLP-1, the single most useful thing you can do is make sure it is written in your chart — the drug, the dose, and who supplies it.

• Tell any clinician, including ones you see for something unrelated. The connection is not obvious and they will not ask.
• Tell anyone scheduling a procedure, early enough that it can change the plan.
• Ask for it to be added to your medication list, not just mentioned. A note in prose does not reach the interaction checker.
• Keep the supplier’s details, because if something goes wrong the pharmacy of record matters — see verify what you received.

The study covers 2021 to 2024, a period defined by shortages that have since closed. Whether the documentation gap has narrowed as the legal basis for compounding changed is not something this data answers — the shifting rules are in which GLP-1s can be compounded.