10 mg diazepam — parent drug and major metabolites
| Compound | Relationship | Approx. half-life | Approx. persistence | Key effects | Very rough amount / exposure from 10 mg diazepam |
|---|
| Diazepam | Parent | ~20–70 h; ~30–50 h is a reasonable central estimate | Parent remains measurable for several days | Anxiolytic, sedative, muscle relaxant, anticonvulsant, amnestic | 10 mg administered |
| Nordazepam / desmethyldiazepam | Major N-demethylated metabolite | ~40–120 h, commonly ~50–100 h | Several days to >1 week | Long-lasting anxiolysis, sedation, muscle relaxation, anticonvulsant effects | ~0.4–4 mg as a deliberately broad modeling range |
| Temazepam | 3-hydroxydiazepam | ~8–20 h | Roughly 1–2 days | Sedative/hypnotic, anxiolytic, muscle relaxation | ~0.2–1 mg as a crude modeling range |
| Oxazepam | Downstream metabolite of nordazepam/temazepam | ~5–15 h | Roughly 1–3 days pharmacologically, although conjugates persist longer | Anxiolytic, sedative, muscle relaxation | ~0.1–1 mg as a crude modeling range |
The major metabolic routes are
diazepam → nordazepam → oxazepam and
diazepam → temazepam → oxazepam. Nordazepam is the major circulating metabolite and has a substantially longer half-life than diazepam.
What I would actually use as a "best guess"
If you want
one set of numbers for a rough conceptual model rather than a range:
| From 10 mg diazepam | Crude working estimate |
|---|
| Diazepam administered | 10 mg |
| Nordazepam | ~1–2 mg |
| Temazepam | ~0.5 mg |
| Oxazepam | ~0.25 mg |
I'd attach
at least a several-fold uncertainty to those metabolite quantities. They should
not be interpreted as measured quantities of metabolites physically generated.
Why? In a human study after a 4-mg dose, urinary recovery over 96 hours was only about
3.9% as nordazepam, 6.6% as temazepam and 2.8% as oxazepam. Those percentages represent
urinary recovery, not total formation, so they cannot simply be treated as the percentage of diazepam converted into each metabolite.
Expected time course after 10 mg
| Time after dose | Most relevant contributors | Directional picture |
|---|
| 0–6 h | Diazepam | Parent drug overwhelmingly important; strongest acute effects |
| 6–24 h | Diazepam + emerging nordazepam | Parent still dominant, but nordazepam is accumulating |
| 1–3 days | Diazepam + nordazepam | Nordazepam increasingly important |
| 3–7 days | Nordazepam + residual diazepam | Long-lived active metabolite becomes increasingly important |
| 7–14+ days | Nordazepam + downstream metabolites | Residual drug/metabolites can remain detectable; this doesn't imply continuous intoxication |
A particularly interesting finding is that after
5 mg, nordazepam's urinary concentration peaked around
100 hours after dosing, compared with about 2 hours for diazepam. The same study found that diazepam and its metabolites remained detectable for at least 15 days.
A useful correction to the earlier 5-mg table
I would now characterize the metabolite amounts more cautiously:
| Dose | Diazepam | Nordazepam | Temazepam | Oxazepam |
|---|
| 5 mg | 5 mg | ~0.2–2 mg | ~0.1–0.5 mg | ~0.05–0.5 mg |
| 10 mg | 10 mg | ~0.4–4 mg | ~0.2–1 mg | ~0.1–1 mg |
These are
modeling ranges, not experimentally measured metabolite yields. The particularly large uncertainty for nordazepam is intentional: about
50–60% of diazepam undergoes N-demethylation, but that does
not mean 50–60% of the administered mass appears as free nordazepam—the metabolite is subsequently distributed, metabolized and conjugated.
Also,
10 mg isn't necessarily exactly twice the subjective effect of 5 mg. Plasma exposure generally increases with dose, but pharmacodynamics, tolerance, age, CYP2C19 genetics, liver function and drug interactions can produce substantial individual differences.
For perspective, controlled human testing has specifically examined
10 mg diazepam and found measurable acute impairment shortly after dosing, while older overnight testing found much less measurable next-morning impairment in healthy subjects. That illustrates why
"half-life" and "duration of noticeable impairment" shouldn't be treated as the same number.