Rybelsus not working? Six things to check before it is the drug

Before you read. This page is a checklist for a conversation, not a substitute for one. It cannot assess whether your treatment is working — that needs your measurements and your prescriber. Nothing here is a reason to change your dose. The label caps intake at one tablet per day and holds each step for 30 days, and climbing faster than that on your own removes the protection those rules exist to give.

Quick answer

Six explanations, in the order worth checking. The first four are more common than the last two, and only the last one is the medicine itself.

It has not been long enough. The first 30 days are on a dose the label says is not effective for glycemic control, and each step then takes four to five weeks to reach steady state.
The morning is costing you the dose. Absolute bioavailability is 0.4% to 1%. A broken window is not a slightly weaker dose; it is a large fraction of a small number.
Something else is in the window. Other tablets, coffee, a full glass of water.
You are on the initiation dose, or the step has not happened. 3 mg is a tolerance rung, not a treatment.
You are measuring something the label does not claim. This tablet is licensed for glycemic control and cardiovascular risk reduction in type 2 diabetes.
The dose is genuinely not enough for you. Which is a prescribing conversation with a ladder above you, not a dead end.

One bit per morning: window kept or broken →

1. It has not been long enough yet

The commonest version of this question arrives in week three, and the label has already answered it.

The starting dose is 3 mg once daily for 30 days, and the dosage section adds, in parentheses, that this dosage is not effective for glycemic control. The first therapeutic dose is 7 mg, starting on day 31. Steady-state exposure is then reached after four to five weeks of daily dosing, because semaglutide's half-life is about a week.

Put those together and the earliest point at which you are on a working dose at a stable level is around day 60. The trials measured their primary endpoint at week 26.

So "three weeks and nothing" is the schedule. "Two months and nothing" is a question. Day 30, day 60, week 26 sets out the whole timeline if you want the arithmetic.

2. The morning is costing you the dose

This is the failure mode unique to an oral GLP-1, and it is invisible from the inside.

Estimated absolute bioavailability of semaglutide tablets is 0.4% to 1%. Absorption happens in the stomach itself, through a narrow window opened by SNAC — an absorption enhancer co-formulated into the tablet. Under one hundredth of what you swallow reaches your bloodstream on a good morning.

Which means the label's four rules are not fussiness. They are the whole margin — every one of them protects a number that has none to spare:

Empty stomach, in the morning
Up to 4 ounces of plain water, and no other liquid
Swallow whole — do not split, crush, chew or dissolve
At least 30 minutes before food, other beverages or other oral medicines

Break one of them most days and you are not on 7 mg. You are on some unknown fraction of 7 mg, and there is no way to feel the difference. Why so little of each tablet gets in covers the mechanism, and what each broken morning costs goes through the individual mistakes.

The trial data on the water is unusually specific and makes the point better than any warning. Doses given with 240 mL of water produced lower absorption than doses given with 50 mL. A full glass is not the careful version of a small one. It is the arm of the study that did worse.

3. Something else is in the window

The 30-minute rule covers food, other drinks and other oral medicines, and the third of those is the one that quietly gets exempted.

People who would never eat inside the window will take their blood pressure tablet, their levothyroxine and their statin alongside it, because those are medicine rather than breakfast. The label makes no such distinction: the wait applies to other oral medications by name.

There is a documented interaction in the other direction too. The label reports that total thyroxine exposure was increased by 33% when levothyroxine was given with semaglutide tablets, and advises monitoring. What the label studied about other medicines works through which drugs were tested and what was found.

If your morning contains other pills, this is worth auditing before anything else on the list, because it is both common and completely fixable.

"Was the morning clean?" is one bit of information a day, and it is the one that settles this question. Takely runs the window from the moment you log the dose and keeps the record, so eight weeks of mornings becomes a column rather than an impression. Open Takely in your browser →

4. You are on the initiation dose, or the step has not happened

Worth checking on the box itself, rather than from memory.

The ladder is 3 mg, then 7 mg from day 31, then 14 mg from day 61 if additional glycemic control is needed. Two things go wrong here more often than they should.

Still on 3 mg past day 30. Sometimes a repeat prescription reissues the starting pack. The label is unambiguous that this strength is not effective for glycemic control, so a month of it beyond the initiation period is a month of tolerance step.

A held step that was never revisited. Steps get held for good reasons — usually gastrointestinal reactions, which the label ties directly to escalation — and a hold is meant to be temporary. It is easy for an intended pause to become the status quo by inertia rather than by anyone deciding.

Neither is something to fix yourself. Both are worth naming at the appointment, because the person prescribing may be assuming a different rung than the one you are on.

5. You are measuring something the label does not claim

This is the uncomfortable one, and it explains a large share of "not working for weight loss" searches.

Rybelsus carries two indications: as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes, and to reduce the risk of major adverse cardiovascular events in adults with type 2 diabetes who are at high risk for them. Weight management is not among them.

Decreased appetite appears in the adverse reactions table — 9% on 14 mg, 6% on 7 mg, against 1% on placebo — which is a description of something the drug does, filed where labels file anything more common on drug than on placebo. It is not an indication, and the label sets no expectation about weight.

None of which means your experience is wrong. It means the yardstick matters: a treatment can be doing exactly what it is licensed to do while the number you are watching sits still. What this tablet is and is not licensed for goes through the distinction, including which semaglutide products do carry a weight indication and at what doses.

6. The dose is genuinely not enough

Sometimes the first five are all clear and the answer is the simple one.

The ladder has a rung above 7 mg for exactly this reason, and the label's own wording — increase to 14 mg if additional glycemic control is needed — describes a decision point rather than an automatic climb. The steady-state numbers show what the step buys: mean concentration of 6.7 nmol/L on 7 mg and 14.6 nmol/L on 14 mg, roughly doubling.

There is also a whole switching framework if the answer is a different product rather than a different number. The labels print conversions between Rybelsus and Ozempic tablets, and between the Wegovy tablet and the Wegovy injection, all at defined equivalences.

What there is not is a version of this you do at home. Taking more than one tablet a day is ruled out by the label in a single line, and the 30-day holds are the mechanism that keeps escalation tolerable.

"It worked and then it stopped"

A slightly different story — and worth separating from everything above.

Two label facts bear on it. First, the ladder is designed so that each step's effect stabilises over four to five weeks — so an effect that seemed to fade a month after starting a dose may be the difference between a rising level and a settled one, rather than a loss of effect.

Second, type 2 diabetes is progressive, and the pooled trial population had been living with it for a mean of 8.8 years at baseline with a mean HbA1c of 8.2%. A treatment holding a number steady in a condition that would otherwise drift is doing something, even when the number is not falling any further.

Both are observations to bring to a prescriber — not conclusions to draw. The reason to know them is that "it stopped working" and "it is holding" can look identical from a single reading and completely different from a series.

What to bring to the appointment

Four items, and they map onto the checklist above.

The dose, and the date it started. Not "about two months" — the date.

The count of clean mornings. Out of the last thirty, how many kept the window. This is the number nobody has and everybody needs; one bit per morning is all it takes to have it.

What else is in the morning. Every other tablet, and whether it goes inside or outside the 30 minutes.

What you are measuring, and what it has done. With dates — so that a change can be lined up against a dose step.

That is the difference between a five-minute conversation that ends in "let's see how you go" and one that ends in a decision.

Frequently asked questions

Why is Rybelsus not working for me?

The commonest explanations are timing and absorption rather than the drug. The first 30 days are on a dose the label says is not effective for glycemic control, each step needs four to five weeks to reach steady state, and under 1% of each tablet is absorbed at all — so a morning routine that breaks the window delivers less than the dose on the box.

Rybelsus 3 mg is not doing anything. Is that normal?

It is what the label describes. The dosage section states that the 3 mg starting dosage is not effective for glycemic control; it is a 30-day tolerance step before the first therapeutic dose.

Can Rybelsus stop working after it has been working?

The label does not describe a loss of effect. Two things can look like one: the four-to-five-week climb to steady state after each dose change, and the progressive nature of type 2 diabetes, where holding a number steady is itself an effect. A single reading cannot tell those apart; a series can.

Should I take two tablets if one is not working?

No. The label states plainly that you should not take more than one tablet per day, and the dose ladder holds each step for 30 days specifically to reduce gastrointestinal reactions. If the dose is not enough, the ladder above you is a prescriber's decision.

Does coffee stop Rybelsus from working?

Coffee inside the 30-minute window reduces absorption, along with food, other drinks and other oral medicines — the label allows plain water only, up to 4 ounces. After the window ends the label places no restriction on coffee at all.

How do I know whether it is the drug or my routine?

Count clean mornings for a month. If most of them kept the window and you are past day 60 on a therapeutic dose, the routine is no longer a plausible explanation, and that is precisely the finding worth taking to an appointment.

Primary sources and further reading

Label text changes between revisions. Every quotation and number above was read off those documents on 8 September 2026.