In clinical practice, lithium is administered by the oral route as tablets or capsules (carbonate) or solution (citrate). Solid pharmaceutical forms of lithium salts include immediate-release and controlled- or extended- or sustained-release forms, depending on the pharmaceutical firm that markets the drugs. Lithium citrate oral solution may be useful in patients unable to swallow capsules or tablets; 5 mL of a commercially available solution contains about 8 mEq of lithium and is approximately equivalent to 300 mg of lithium carbonate. The relatively high water solubility of lithium salts used in therapy permits rapid and total solubility of solid pharmaceutical forms containing lithium salts within the gastrointestinal tract fluids, leading to good oral absorption of lithium – in terms of the amount absorbed by the gut. Moreover, the rate of oral absorption of lithium salts depends on the rate of solubility of the different lithium salts and the pharmaceutical forms, i.e. immediate- or sustained-release forms, within the intestine.

Following oral ingestion, fast and effective absorption is observed within 1–6 h for the immediate-release form (later for the sustained-release form) through the upper gastrointestinal tract (Malhi and Tanious 2011). A complete dissociation of lithium salts occurs after their oral absorption, giving rise to lithium as a single cation in the circulating blood. Chloride and sulphate salts are rapidly and almost completely absorbed, while the less water-soluble carbonate salt is absorbed more slowly (Lehmann 1997). Lithium is mainly absorbed in the jejunum and ileum (Bettinger and Crismon 2006). Although the mechanisms involved are not well known, lithium transport across the intestine involves both active and passive mechanisms (Kersten et al. 1986). The bioavailability of all formulations of lithium is good, with a range of 80–100 % (Hunter 1988; Price and Heninger 1994). Some gastrointestinal factors influence the oral absorption of lithium; for example, food may influence the rate and extent of absorption. Studies by Gai and colleagues (1999, 2000) using two different sustained-release formulations of lithium showed that the extent (bioavailability) of oral absorption was not modified under fasting or fed conditions. However, meals (normal diet, high fat and high fat/protein meals) significantly increased the maximal serum concentration (C _max) of lithium as compared with fasting conditions, suggesting that the rate of lithium oral absorption may be increased by meals, even if the observed time of maximum plasma concentration (T _max) was similar across fed and fasting conditions (Gai et al. 1999, 2000).

Because of their physicochemical properties and small molecular size, lithium ions largely distribute into extracellular and intracellular body water spaces, as do sodium ions. Among all the models tested, the best model to describe lithium serum PK is a two-compartment model with first-order absorption rate constant (ElDesoky et al. 2008; Findling et al. 2010; Sproule et al. 2000; Taright et al. 1994). However, as shown in Table 2.1, some studies have also used a one-compartment model (Hoegberg et al. 2012; Yukawa et al. 1993) or a non-compartmental analysis. However, a lack of robustness of the analytical methods used, and the number of subjects and time points, did not always permit accurate determination of the number of compartments. Lithium, like sodium ion, is water soluble and its binding to plasma proteins is negligible (Clarke et al. 2004; Price and Heninger 1994). Unlike other organic drugs used in mental disorders, lithium has a slow diffusion from extracellular to intracellular compartments due to its low passive diffusion permeability across cell membranes. The means of the central volume of distribution of lithium calculated in several studies were from 0.8 to 1.2 L/kg with an initial distribution into the total extracellular body fluids (Ward et al. 1994). Therefore, lithium has a small volume of distribution, slightly higher than the total body water volume, which makes it susceptible to large variations of total body water volumes that are frequently observed following some pathophysiological changes. For example, with ageing, alteration of body composition such as increased body fat, decreased fat-free mass and total body water will decrease the volume of distribution of lithium, decreasing the time to reach the steady state with repeated doses of lithium together with a reduced elimination half-life and an increase of the C _max/C _min ratio. Oedemas or, conversely, dehydration syndrome may also change the percentage of total body water and consequently the volume of distribution of lithium. The distribution of lithium into the tissue from the blood occurs variably at differing rates: for example, high and low uptake rates occur in the liver/kidneys and brain/muscles, respectively. The kidney, bone and thyroid lithium concentrations are twice those simultaneously measured in serum (Amdisen 1977; Kusalic and Engelsmann 1999; Wilting et al. 2007) with a T _max from 2 to 4 h later than that observed in the serum. Lithium concentrations in the liver and muscles are lower than those in serum and other tissues (kidney, thyroid, bone). Erythrocyte concentrations of lithium correlate in most patients with plasma levels but are subject to large interindividual variability. In early studies, lithium concentration in RBC was proposed as a better surrogate of brain lithium concentration than serum lithium concentration. Today, this statement remains poorly substantiated, and lithium RBC concentrations are not consistently used in clinical practice (Lee et al. 1975; Schreiner et al. 1979; White et al. 1979).

Lithium is not subject to first-pass hepatic metabolism (Bauer et al. 2006). A small amount (2 %) of lithium is eliminated in the bile within 24 h of its oral administration (Terhaag et al. 1978). Thus, the PK of lithium is not influenced by any changes in hepatic biotransformation or hepatic blood flow.

Lithium is freely filtered by the renal glomerulus with a filtration rate similar to that of other small molecules circulating freely in the plasma that are unbound to proteins. Once filtered, about 75–80 % of lithium is actively reabsorbed in the proximal tubules, while a small proportion can also be reabsorbed in loop of Henle and by the epithelial channel which involves sodium (eNaC) in the early distal tubules (Hayslett and Kashgarian 1979) under sodium-restricted conditions (Thomsen and Shirley 2006; Trepiccione and Christensen 2010). The rate of lithium renal elimination is linearly correlated with its serum levels (Bauer et al. 2006). The terminal elimination half-life (t _1/2) of lithium in the plasma ranges from 16 to 30 h in healthy volunteers and patients with normal renal function (Arancibia et al. 1986; Cooper et al. 1978; Granneman et al. 1996; Hunter 1988; Lee et al. 1998; Luisier et al. 1987; Thornhill 1978). This elimination phase can be observed as early as 4–6 h after the T _max. The lithium elimination half-life during long-term lithium treatment or in patients with impaired renal function increases and can reach 50 h. A linear relationship between the dose and the AUC of serum lithium levels versus time has been determined following several oral doses of lithium (Amdisen 1977); this linear relationship is an indicator of any saturation of the different PK phases (absorption, distribution, elimination) of lithium at usual therapeutic doses (12–36 mmol/24 h). However, this does not argue against the involvement of carrier-mediated processes in absorption, distribution and excretion, as sufficient concentrations to saturate and kinetically reveal these transport systems could not easily be reached in vivo.

Lithium is mainly eliminated by the kidneys as a free ion. Its clearance ranges from 0.6 to 2.4 L/h (Ward et al. 1994) with large interindividual variability (Amdisen 1977; Frye et al. 1998; Perry et al. 1981). This variability has been identified in PK population studies (Taright et al. 1994; Yukawa et al. 1993), interindividual lithium clearance variability being from 25 % to 38 % in bipolar patients with normal renal function. Lithium total body clearance has been reported as variable depending on several pathophysiological states, such as age, total body weight and renal function (ElDesoky et al. 2008). It has been clearly demonstrated that the effect of weight parameters (total body weight or lean body weight) and creatinine renal clearance accurately predict steady-state lithium concentrations by decreasing interindividual variability (Sproule et al. 2000). Renal lithium clearance ranges from 10 to 40 mL/min and decreases with age (linearly with age-related decrease in renal glomerular filtration), low sodium intake, volume depletion in dehydration or with age, cardiac failure, concomitant use of drugs affecting renal glomerular filtration such as nonsteroidal anti-inflammatory agents (see Sect. 2.5.1.3), kidney disease and any event increasing proximal tubular reabsorption of sodium and water (i.e. dehydration, decreased salt intake, extra-renal salt loss, diuretics) and increased with pregnancy and total body weight.

In the case of significant impaired renal function, plasma lithium concentrations increase quickly and may trigger acute toxicity symptoms. A preliminary check of renal function is thus highly recommended before initiation of lithium treatment, followed by regular checks throughout the course of treatment (see Sect. 2.8). Lithium by itself causes a modest decline in renal function, thus increasing its elimination half-life and steady-state levels. End-stage renal disease (ESRD) is a very rare complication of long-term lithium treatment affecting only 1 % of patients over 15 years of treatment (Tredget et al. 2010). No study has evaluated the variability of the PK parameters for patients with ESRD. A known renal failure after lithium initiation is considered a contraindication, especially when a sodium-poor diet is also prescribed (Grandjean and Aubry 2009a).

Reiss and colleagues (1994) found that, in obese patients, the steady-state volume of distribution in litres was not significantly different from that in normal weight individuals (44.5 ± 9.8 L versus 41.3 ± 8.4 L in obese and nonobese subjects, respectively), which is consistent with the similarly calculated ideal body weight (IBW), and thus significantly less as expressed in litres/kg (0.42 ± 0.09 L/kg and 0.66 ± 0.16 L/kg for obese and nonobese patients, respectively). Total lithium clearance was greater in obese patients (33.9 ± 7.0 mL/min in obese patients versus 23.0 ± 6.2 mL/min in nonobese subjects). The authors thus demonstrated that the volume of distribution of lithium was significantly correlated with ideal body weight and fat-free mass but not with total body weight. Lithium clearance was significantly correlated with total body weight but not with creatinine clearance. The major conclusion was the necessity to increase the maintenance dose in obese patients to achieve similar lithium steady-state levels as those observed in nonobese patients, because of a higher lithium clearance in obese patients. However, the starting dose should be the same as that used in nonobese patients since the volume of distribution of lithium in obese patients is more related to the ideal body weight than the total body weight.

Circadian rhythm has a specific influence on the PK of lithium. A study in rats by Olesen and Thomsen (1985) reported a significant decrease in serum lithium concentration during the early night portion of the light-dark cycle, when rats are most active, associated with a large increase in renal clearance of lithium. In mice, the toxicity of acute intraperitoneal injection of lithium chloride (LiC1) (940 mg/kg) increased from 20 % during the night to 70 % during the day (Hawkins et al. 1978). Similarly, Shito and colleagues (1992) reported time-dependent changes in both lithium clearance and toxicity. Chapter 6 is devoted entirely to this topic.

Elderly bipolar disorder patients require a lower dose of lithium to achieve the same steady-state levels observed in young adults with normal renal clearance (see Sect. 2.6.2). In contrast, child bipolar disorder patients have a shorter elimination half-life and greater clearance compared to adults (see Sect. 2.7.1).

Associated with age, hypertension and congestive heart failure may decrease the glomerular filtration rate and accordingly further reduce lithium renal excretion (Sproule et al. 2000), indicating that a lower dose of lithium may be required in these patients.

Briefly, lithium clearance increases by 30–50 % in pregnant women due to an increase in its renal glomerular filtration rate, especially during the last months of pregnancy. There is a rapid return to the pre-pregnancy state immediately after delivery (Grandjean and Aubry 2009b; Schou 1990; Viguera et al. 2002). Therefore, TDM must be carried out more frequently for pregnant women. Lithium also diffuses freely from maternal plasma to breast milk (Schou 1990; Yoshida et al. 1999) with serum concentrations in infants ranging from 10 % to 50 % (and possibly up to 80 %) of those in maternal plasma (Viguera et al. 2002). All studies in pregnancy and breastfeeding populations are necessarily retrospective, with the safety of the infant-mother couple being the first concern. (See Chap. 18.)