He started at about 0.8 grams a day and did beautifully. Then he climbed, and everything he had gained came back off him in the order it had arrived.
A twenty-five-year-old with Tourette's and OCD started at about 0.8 grams a day. His tics settled. He came alive. He started teaching a class of new DJs. Then exams arrived, his OCD flared, he smoked more, and he began a climb he told nobody about: 0.8 grams to five a day, then to ten. Anxiety. Brain fog. Motivation gone. Tics roaring back worse than before. Finally a severe psychosis and a hospitalisation.
It had worked at low doses. It always worked again at low doses. But he kept climbing, and eventually the administration gave up on him. He is not permitted cannabis at all now.
That is the biphasic curve, and it is the whole answer to the question in the title. More of a good thing is not more good. Past a certain point it is the opposite thing, wearing the same label, and the way it announces itself is by looking exactly like a supply problem.
For my first decade I coached with a model I had inherited rather than examined: more of a good thing is more good. When someone told me the medicine had stopped working, I heard a supply problem. Stronger flower. Bigger gummy. I was confident, and I was wrong in a way that took years to find out about.
What changed my mind was not a paper. It was noticing that the patients who did best were never on the biggest amounts. They had found a small amount that worked and they defended it, sometimes against me.
It is a position on a curve. Two people take the identical number and land on opposite sides of the same molecule's behaviour. One calmer, one worse. The shape is a hill, and the cruel part is that you come down the far side while still taking more.
There is one study I would put in front of anyone who doubts this. Childs, Lutz and de Wit, 2017. Forty-two healthy adults, within-subjects, so everyone served as their own control. Three conditions: 0 mg, 7.5 mg, or 12.5 mg of oral THC, then the Trier Social Stress Test. At 7.5 mg, THC reduced subjective distress. At 12.5 mg, it increased negative mood and impaired performance. Not helped less. Worse than placebo. ESTABLISHED
Two doses. Five milligrams apart. Opposite directions.
Five milligrams is a bite off the corner of a 10 mg gummy. It is the difference between a patient who measured and a patient who was hungry.
I will give you the other half, because a page that only argues one side is not worth reading. That THC's anxiety effects are highly dose-dependent is SUPPORTED (Sharpe 2020). That "cannabis is biphasic" is a proven law is OVERSTATED. The curve is a good working model. It is not a certainty, and anyone selling it to you as one is selling.
If you want somewhere to put your own numbers while you read the rest of this, the free WIZDOM journal is the tool built for exactly that, and it costs nothing.
Your window is the range between the amount that is too little to do anything and the amount that starts working against you. Below it, nothing. Inside it, the thing you were hoping for. Above it, the anxiety, the fog, the symptom coming back louder.
Two facts sit under the old advice to start low and go slow, and both of them are about the window. The window is narrow. And individuals vary so much that no population-level safe dose can be set SUPPORTED (Kitdumrongthum 2023). That is not a hedge. It is the reason nobody, me included, can tell you your number from a web page.
Titration, in plain words: starting low and going up slowly, a little at a time, until you find the amount that is just right. Published practice guidance for a first step exists, it is SUPPORTED, and the number in it is far smaller than almost any patient expects. I am deliberately not printing it here, because a number on a page becomes an instruction the moment somebody reads it. Ask your prescriber for the starting point, and ask them what waiting for full onset means for the route you are using.
And here is the honest gap. No titration trial exists. That titrating improves outcomes is HYPOTHESISED. I believe it, I have watched it work for twenty years, and it is still hypothesised. I would rather tell you that myself than have you find out later.
Everything in a patient's experience of medicine teaches them that a bigger number is a stronger effect. Paracetamol is prescribed at 500 mg, two or three times a day, very specifically. Cannabis is prescribed up to a quantity per month, which is not the same kind of instruction at all, and almost nobody explains the difference.
So when relief fades, the reasonable move looks obvious. Take more. And if you were already near the top of your curve, taking more moves you past it, which makes things worse, which reads as further proof that you need more still. That is the loop. It runs quietly for months, it is invisible from the inside, and it is obvious from the outside, which is the entire argument for not doing this alone.
At the chronic top end sits cannabinoid hyperemesis syndrome, where the antiemetic becomes the emetic. In Sorensen 2017, 96.8% of cases resolved on cessation ESTABLISHED. The mechanism is UNKNOWN. I am not going to invent one for you.
Fading relief has more than one explanation and they point in different directions. You may be over the top of your curve. Or the peak may have moved under you, which is tolerance, and receptor recovery is a real and reversible thing with its own timeline. I have written about that separately, because the timeline is shorter than the folklore says: see the tolerance break guide and why medical cannabis stops working.
The point for this page is narrower. Those two explanations call for opposite responses, and only one of them is made worse by guessing. So the sentence to hold on to is this: a fading effect is a question to investigate, not a number to raise.
I explain and monitor pharmacology. I am not a prescriber and I do not act like one. Anything attached to a lab value or a prescription goes to the prescribing doctor, every time.
That boundary does not move. Hard edges make you forecastable, and forecastable is the currency of working beside a medical team. It is also why this page contains no protocol for you. It is education about a shape, so that you can have a better conversation with the person who is actually allowed to change your numbers.
You find it the boring way. One variable at a time, with the prescriber in the loop, and written down. What you took, when, by what route, and what happened over the following hours. Memory is not evidence. Most patients who tell me the medicine stopped working changed three things at once and cannot say which one mattered, and the log, not the recollection, is what makes the pattern visible. That is the whole reason the journal exists and the whole reason it is free.
Your job, and mine, is not to talk anyone out of their medicine. It is to find the amount at which it is still medicine, and then help you hold that line.
Biphasic means the same substance produces opposite effects at different amounts. The shape is a hill. Going up helps, and then, past a turning point, going further up starts to hurt, while you are still taking more. In Childs, Lutz and de Wit 2017, 7.5 mg of oral THC reduced distress under a laboratory stress test and 12.5 mg increased negative mood and impaired performance.
Because you have gone past the top of the curve. Below the turning point more THC does more of what you wanted. Above it, more THC does the opposite, and the anxiety, the fog and the returning symptom read as proof that you need even more. That reading is the trap, and it is the reason people climb for months in the wrong direction.
It is the range between the amount that is too little to do anything and the amount that starts working against you. Two things are known about it: the window is narrow, and individuals vary so much that no population-level safe dose can be set (Kitdumrongthum 2023). That is why nobody can tell you your window from a page. You find it with your prescriber, slowly, and you write it down.
Take it as a clue, not a verdict, and take the question to your prescriber before you take anything else. Fading relief has several possible explanations, including tolerance and including the fact that you are already over the top of your curve, and those two point in opposite directions. Increasing is the one move that makes the second explanation worse.
No, and I will not sell it to you as a law. That THC's anxiety effects are highly dose-dependent is SUPPORTED (Sharpe 2020). That cannabis is biphasic as a proven universal rule is OVERSTATED. The curve is a good working model and a bad certainty, which is exactly why the answer for any one person comes from observation rather than from a table.
At the chronic top end sits cannabinoid hyperemesis syndrome, where the antiemetic becomes the emetic. In Sorensen 2017, 96.8% of cases resolved on cessation, which is ESTABLISHED. Why it happens at all is UNKNOWN. Cyclical vomiting in a long-term heavy user is a prescriber conversation, not a dosing adjustment.
Please read this part. This is patient education, not medical advice, and I am not a physician and never a prescriber. It is written from twenty years of education work and patient coaching, not from a controlled trial, and it is not a dosing protocol. Cannabis treats nothing; patients report outcomes. Do not change a prescribed regimen, cannabis or anything else, without talking to your prescriber, particularly if you take other medications. If something is getting worse, that conversation should happen now rather than after another experiment.
The book is free. It is called The Coach Who Wasn't a Doctor, every load-bearing claim in it carries its evidence tier, and the chapters I had to rewrite after I refuted myself are still in there.
Read the free book