The saliva test does not detect CBD. It looks for THC. And that is precisely why a CBD user can test positive: the flowers and resins authorised in France contain traces of THC, up to 0.3%, and those traces are real THC.
Second point to keep in mind straight away: in France, there is no tolerated threshold. If the laboratory analysis confirms the presence of THC, the offence of driving after using narcotics is made out, whether that THC comes from a cannabis joint or from a legally purchased hemp flower.
This guide explains what the test measures, why it can react, how a roadside check goes and what you can ask for. It gives no method for beating a screening test: no reliable one exists, and that is not the subject. Information up to date as of September 2026.
Key points — Roadside saliva screening targets THC, never CBD. A compliant CBD flower or resin (0.3% Δ9-THC maximum) contains traces of THC that can turn the test positive, especially after inhaling. The roadside test is only a screening: it is the laboratory analysis that counts as evidence, with a detection threshold of 1 ng/ml of saliva. You can ask for a blood sample to reserve your right to a second expert analysis. Detection times vary and none is guaranteed.
What a saliva test detects (and what it does not)
The test used at the roadside is an immunoassay. A collector picks up a little saliva, which travels along a strip containing antibodies. Each antibody is designed to recognise one family of molecules. If the target molecule is present above the test's sensitivity, a line appears or disappears depending on the model.
The families looked for are set by regulation: cannabis (via THC), cocaine, opiates, amphetamines. CBD is not one of them. It is not classed as a narcotic and no antibody in the test is meant for it. The same goes for CBG and CBN.
The thresholds set by the arrêté of 13 December 2016
This arrêté (ministerial order) imposes minimum performance levels:
- the screening test must detect THC from 15 ng/ml of saliva;
- the confirmatory laboratory analysis must detect it from 1 ng/ml of saliva;
- the blood analysis, from 0.5 ng/ml of blood.
Beware of the misreading: these are not permitted levels. The Cour de cassation (France's highest criminal court) ruled on this on 21 June 2023: they are detection thresholds, not thresholds for the offence. As soon as the laboratory establishes the presence of THC, the offence exists, whatever the dose.
Why a CBD user can test positive
The permitted traces of THC
One gram of flower at 0.3% Δ9-THC contains up to 3 mg of THC. CBD flowers, pre-rolls and resins are full-spectrum products by nature: plant material, raw or concentrated, with all its cannabinoids. The CBD percentage on the label tells you nothing on this point. A Critical Mango at 21.01% CBD or a Banana Dry, an ice-water hash at 36.93%, are subject to the same THC ceiling and can contain traces of it.
Your mouth is on the front line
The test does not analyse your blood, it analyses your saliva. When you smoke or vaporise, THC is deposited directly on the lining of your mouth. That is the main source of THC in saliva in the hours that follow. Combustion, incidentally, is harmful to your health whatever the plant: see the risks and the alternatives.
Repeated use
THC is lipophilic: it is stored in fat and eliminated slowly. Daily use, even of products very low in THC, does not have the same profile as a one-off use.
A product that does not match its label
This is the other scenario. A flower advertised as "less than 0.3%" can exceed the limit if the seller checks nothing. The only document that lets you verify is the batch's certificate of analysis, issued by an independent laboratory. Demand it from any seller. To read it: decoding a COA.
What the studies say, and their limits
Italian researchers measured THC and CBD in the blood, saliva and urine of volunteers after they consumed "light cannabis", the local equivalent of CBD flowers (Pacifici et al., 2020). They conclude that saliva is a relevant matrix for monitoring this kind of use, which amounts to saying that THC can be measured in it. These were small samples and specific products: you cannot draw from them a percentage of risk that applies to your case.
How long afterwards? Why nobody can give you a number
This is the most frequently asked question, and the honest answer is disappointing: it depends, and nobody can guarantee you a waiting time.
What makes the detection time vary:
- the amount consumed and the product's actual THC content;
- the way you consume it, with inhaling loading the saliva far more;
- the frequency: occasional or daily;
- your metabolism, your body fat, your hydration;
- the sensitivity of the method: a laboratory analysis at 1 ng/ml sees things that a field test at 15 ng/ml does not.
An order of magnitude from the literature: with THC-rich cannabis vaporised by occasional users, THC in saliva peaks about ten minutes after use and then declines rapidly over the following hours (Arkell et al., 2019). With a hemp flower, the starting quantities are far lower. But these data come from controlled trials with a single dose. They do not describe a regular user, and they are not a calculation rule.
The charts announcing "6 hours for this, 24 hours for that" give a false impression of precision. None has any legal value. If you have to drive, the only certainty is not to have consumed any product containing THC.
How a roadside check unfolds
Step 1 — Roadside screening
The police can carry out a screening test after an accident, after a traffic offence, if there are reasons to suspect the use of narcotics, but also during organised checks with no particular reason. You rub a collector around your mouth, and the result shows within a few minutes. Refusing to take the test is a separate offence, provided for by article L235-3 of the Code de la route (French Highway Code): it is not a way out.
Step 2 — If the test is positive, the sample
A second saliva sample is taken to be sent to a laboratory. At that point, the officer must ask you whether you wish to reserve the option of requesting a technical examination or a second expert analysis (article R235-6). If you answer yes, a blood sample is taken as soon as possible. At the same time, the positive screening in practice leads to the police holding on to your licence.
Step 3 — The laboratory analysis
This is the one that counts. The laboratory looks for THC with a far more precise method than the test strip. If the analysis is negative, the procedure stops. If it is positive, it establishes the use of narcotics within the meaning of article L235-1.
Step 4 — Notification and the second expert analysis
Once you have been notified of the results, you have five days to ask the public prosecutor, the investigating judge or the trial court for the second expert analysis (article R235-11), provided you reserved that option at step 2. Within the same time limit, you can also ask for a search for psychoactive medicines.
What to do if you test positive after CBD
- Stay calm and cooperate. A refusal or aggressive behaviour makes things worse.
- Answer yes to the question about the second expert analysis. Without a blood sample at that moment, you lose that right.
- Note the time, the place, how things went, what you were told. The procedure is tightly regulated, and irregularities in it are a classic line of defence.
- Gather the invoice, the packaging and the product's certificate of analysis. Be clear-eyed about what they are worth: since the 2023 ruling, the legal origin of the THC does not make the offence go away. These documents serve to explain the context, not to secure an automatic acquittal.
- See a lawyer specialising in traffic law, quickly, because of the five-day time limit.
The penalties you face (three years' imprisonment and a €9,000 fine since the law of 9 July 2025, loss of half your points) are set out in CBD and driving in 2026.
False positives and false negatives: how reliable field tests are
Roadside tests are not perfect. An evaluation of two commercial devices, published in 2019 (Arkell et al., Drug Testing and Analysis), found false negatives and false positives on both, with neither reaching the recommended performance levels on every criterion. That is why the law requires laboratory confirmation.
Two practical consequences. A negative screening does not prove the absence of THC. And a positive screening that you believe is unjustified is challenged through the analysis and the second expert analysis, not by arguing on the spot.
What about tests away from the road?
In the workplace, in occupational medicine or in sport, other tests exist, often urine tests. They look for a THC metabolite that stays detectable far longer than THC in saliva. The rules, thresholds and consequences have nothing to do with the Code de la route. If this concerns you, find out about the precise framework that applies to you.
For the general framework for CBD in France, see what the law says in 2026.
Sources
- Arrêté du 13 décembre 2016 fixant les modalités du dépistage des substances témoignant de l'usage de stupéfiants, articles 3 and 10.
- Code de la route, article L235-1; articles L235-2 and L235-3; articles R235-6 and R235-11.
- Cour de cassation, chambre criminelle, 21 juin 2023, pourvoi n° 22-85.530.
- Service-Public, « Conduire après avoir consommé du CBD est interdit », 31 July 2023.
- MILDECA, « Le CBD », page updated on 21 May 2025.
- Arkell TR et al., « Detection of Δ9 THC in oral fluid following vaporized cannabis with varied cannabidiol (CBD) content: An evaluation of two point-of-collection testing devices », Drug Testing and Analysis, 2019.
- Pacifici R, Pichini S, Pellegrini M et al., « THC and CBD concentrations in blood, oral fluid and urine following a single and repeated administration of "light cannabis" », Clinical Chemistry and Laboratory Medicine, 2020.
