Lithium
Lithium. Monovalent cation — exact mechanism unclear; inhibits GSK-3β, IMPase (depleting myo-inositol), and modulates neurotransmitter release.… Oral half-life 15 h.
Monovalent cation — exact mechanism unclear; inhibits GSK-3β, IMPase (depleting myo-inositol), and modulates neurotransmitter release. First-line mood stabilizer for bipolar maintenance + suicidality reduction. Narrow therapeutic index (0.6–1.2 mEq/L) — TDM mandatory. Renal-cleared; thiazides + ACEi raise levels.
PK: Valecha 1990 (PMID:2079355), 60 patients, single 900 mg lithium carbonate, verbatim "volume of distribution 0.62 +/- 0.26 l/kg" (x 70 kg = 43.4 L) and "serum half life 15.34 +/- 6.06 h". The previous citation (PMID:10890582) was a food-effect study of a sustained-release matrix that quantifies none of these. ka and F dropped 2026-09-06 to solver defaults: both are supportable from Thornhill 1978 (PMID:365541) — absorption t1/2 0.78 h for the ordinary form, so ka 0.89/h, and "the ordinary preparation was completely absorbed" — but that is a second study and the F claim is qualitative. CAVEATS: (1) half-life is basis-dependent, 15.34 h single-dose here vs 28.9 h after chronic dosing (PMID:657687), and this record stores the single-dose value because the solver superimposes intake events; (2) six of seven model-selection papers fit TWO compartments with a redistribution half-life around 4-5 h, so the first several hours of this one-compartment curve are approximate; (3) renal lithium clearance falls with sodium depletion, dehydration and reduced GFR (PMID:6380872), so a fixed-clearance curve understates toxicity risk in those states; (4) the alias Lithobid is a sustained-release product, for which absorption is ~85%, not complete. PK: A THERAPEUTIC DOSE RENDERS AS A TOXIC SERUM LEVEL, AND THE CAUSE IS A CATALOGUE-WIDE CONVENTION RATHER THAN A WRONG NUMBER. Every stored value is verbatim and correctly sourced. But the dose is nine hundred milligrams of lithium CARBONATE while the volume, the half-life and the reported serum concentration all describe the lithium ION, and only 18.8 percent of that salt mass IS lithium — so the solver divides a salt mass by an ion volume and OVERSTATES THE PLASMA CURVE 5.32-FOLD. In the units this drug is actually monitored in, the model renders a single standard dose at 2.32 mmol/L against a therapeutic band of 0.6 to 1.2 and a toxicity threshold around 1.5; with the ion mass it renders 0.44, which is the right order. THE SECOND DEFECT PULLS THE OTHER WAY AND IS WHY NEITHER IS FIXED ALONE: no bioavailability is stored, so the solver substitutes 0.9 against a volume that is itself an APPARENT one from an oral study, dividing by absorption twice — pinning it at unity is correct and would make the toxic reading WORSE by a further ten percent. Both move together or neither moves. AND THE CONVENTION IS NOT LITHIUM-SPECIFIC: this catalogue deliberately stores free-base masses beside salt-form doses, which is harmless while the mass is only used to convert micromolar affinities and is NOT harmless once the dose feeds the plasma curve. Brompheniramine has the same shape at 1.36-fold, at least twenty-nine records discuss a salt-versus-base distinction in their prose, and lithium is merely the extreme because its counter-ion is most of the molecule. LOGGED AS A SCHEMA GAP: there is no field that says which moiety a dose is expressed in. DOSE MOIETY: THE RECORD WAS RENDERING A THERAPEUTIC DOSE AS A TOXIC SERUM LEVEL, AND NOT ONE OF ITS NUMBERS WAS WRONG. Every stored value is verbatim; the defect was that the dose is nine hundred milligrams of lithium CARBONATE while the volume, the half-life and the concentration this drug is monitored by all describe the lithium ION, of which the salt is only 18.787 percent. The model asserted 2.32 mmol/L for a standard dose against a therapeutic band of 0.6 to 1.2 and a toxicity threshold near 1.5. WITH THE MOIETY DECLARED IT RENDERS ABOUT 0.49, which is the right order for a single dose. THE BIOAVAILABILITY MOVES IN THE SAME COMMIT AND DELIBERATELY SO: it was absent, so the solver substituted 0.9 against a volume that is itself an APPARENT one from an oral study, dividing by absorption twice — and pinning it at unity, which is the correct convention and is supported by a second study reporting that "the ordinary preparation was completely absorbed", makes the uncorrected reading ten percent WORSE. Fixing either alone would have been the atracurium mistake. The displayed dose is unchanged: a nine hundred milligram tablet is still logged and shown as nine hundred milligrams.
Pharmacokinetics
- Oral — t½ 15.3 h · F 100% · Vd 43.4 L
Sources: PMID:365541, PMID:2079355, PMID:9169962, PMID:657687, PMID:6380872