Lithium levels and older adults

Older patients will usually require lower doses of lithium to achieve therapeutic blood levels, owing to age-related pharmacokinetic and pharmacodynamic changes, polypharmacy, drug–drug interactions and comorbidities, which may increase the risk of lithium toxicity. ^{Reference Sun, Herrmann and Shulman5} The current reference range for lithium in UK laboratories is 0.4–1.0 mmol/L, with some variations globally. ^{Reference Shulman, Almeida, Herrmann, Schaffer, Strejilevich and Paternoster4} National Institute for Health and Care Excellence (NICE) guidelines recommend maintaining plasma lithium levels between 0.6 and 0.8 mmol/L for first-time users and between 0.8 and 1.0 mmol/L for those who relapse or have subthreshold symptoms. ¹ However, these guidelines do not specify lithium level monitoring in an age-specific way. Lithium toxicity, which often manifests as neurotoxicity (87%), nephrotoxicity (47%) and cardiovascular toxicity (45%), can become a serious concern, especially in older patients, many of whom receive polypharmacy. ^{Reference Sun, Herrmann and Shulman5} Sun et al, in their systematic review of case reports of lithium toxicity in older people, suggest that significant adverse effects or lithium toxicity may occur at levels of over 0.8 mmol/L in over-75-year-olds. ^{Reference Sun, Herrmann and Shulman5}

Adverse effects: the renal system

Since the renal system almost entirely clears lithium, reduced renal function, which is common in older people, increases the risk of lithium accumulation, necessitating lower oral doses. Lithium-induced nephrogenic diabetes insipidus can cause polyuria, secondary thirst and sometimes irreversible damage. Although the condition may be reversible within the first 2–6 years, about 20% of patients develop irreversible diabetes insipidus. Lithium has also been implicated in renal damage seen in pathological samples, such as cortical scarring and microcysts, collectively referred to as lithium-induced nephropathy. ^{Reference Schoretsanitis, Filippis, Brady, Homan, Suppes and Kane6}

A systematic review and meta-analysis found that about 25% of lithium-treated patients developed chronic kidney disease (CKD), a significantly higher prevalence than in the general population (at around 5%). ^{Reference Schoretsanitis, Filippis, Brady, Homan, Suppes and Kane6} In patients aged 65 and older, this prevalence increased to one-third of individuals. The risk of CKD is more pronounced with prolonged lithium use, with a two-fold higher risk compared with non-lithium treatments. However, factors such as body mass index, smoking, hypertension and diabetes did not significantly influence CKD risk, although cardiovascular disease showed a near-significant association. ^{Reference Schoretsanitis, Filippis, Brady, Homan, Suppes and Kane6} Although there is some controversy regarding the extent of CKD risk associated with lithium versus other treatments, ^{Reference Bosi, Clase, Ceriani, Sjölander, Fu and Runesson7} older people appear to be at a higher risk. Individuals with higher levels of lithium (above 0.8 or 1.0 mmol/L) were at higher risk of acute kidney injury. ^{Reference Bosi, Clase, Ceriani, Sjölander, Fu and Runesson7} Rej et al demonstrated in a population cohort study that lithium levels above 0.7 mmol/L increase the risk of renal decline, whereas levels below or equal to 0.7 mmol/L do not lead to this risk in older populations. ^{Reference Rej, Herrmann, Gruneir, McArthur, Jeyakumar and Muanda8}

Safe and effective dosing

The question of whether, even at lower levels, lithium is safe (from a renal perspective) in older people without affective disorder has been addressed to some extent. Aprahamian and colleagues examined renal safety in older people receiving the drug for mild cognitive impairment in a placebo-controlled trial. In this study, there were no adverse renal effects with lithium use at low doses (150–600 mg daily) over 2 years. ^{Reference Aprahamian, Santos, Dos Santos, Talib, Diniz and Radanovic9} The target lithium range in this study was 0.25–0.5 mmol/L, lower than what is typically used for bipolar disorder.

There is also evidence from a cohort study ^{Reference Coppen and Abou-Saleh2} and a randomised controlled trial ^{Reference Coppen, Abou-Saleh, Milln, Bailey and Wood10} in people with unipolar or bipolar disorder that dosage reduction (by 25–50%) (lithium levels between 0.45 and 0.79 mmol/L) ^{Reference Coppen, Abou-Saleh, Milln, Bailey and Wood10} significantly reduced affective morbidity. In addition, thyroid-stimulating hormone levels were decreased in this group, as were total side-effects and tremor. ^{Reference Coppen, Abou-Saleh, Milln, Bailey and Wood10}

Two Delphi studies have addressed the issue of whether lower levels of lithium meet the right balance of being both safe and effective in older adults with bipolar disorder. An international group of experts for the International Society for Bipolar Disorder (ISBD) reached consensus on using plasma monitoring of 0.4–0.8 mmol/L for those aged 60–79 and 0.4–0.7 mmol/L for those over 80. ^{Reference Shulman, Almeida, Herrmann, Schaffer, Strejilevich and Paternoster4} A second Delphi study of old age psychiatrists in Germany showed that there was consensus on slightly lower plasma levels, of 0.4–0.7 mmol/L for patients under 80 and 0.4–0.6 mmol/L for those aged 80 and over. ^{Reference Christl, Müller-Oerlinghausen, Bauer, Kamp, Fußer and Benninghoff11} Experts in both studies agree that maintaining lower lithium concentrations is effective and safer for older patients, given concerns about kidney function and neurotoxicity. A group of us in the UK, from the Faculty of Old Age Psychiatry of the Royal College of Psychiatrists, prepared a talk on this topic for the Faculty’s annual conference held in Liverpool in 2025, presenting this evidence base. We surveyed the audience of predominantly old age psychiatrists, trainees and staff grades (93 in in total), who voted after the talk. The highest consensus was on the question asking whether we should recommend change to lithium monitoring for people over 65 years old in laboratories in the UK and 82% agreed, 9% disagreed and 9% were not sure. In a further question about whether they agreed with the ISBD guidance, 84% agreed, 3% disagreed and 8% were unsure. So, the largely UK audience supported by a large majority that current lithium plasma monitoring advice needs to be changed.

Implications for clinical practice

In the UK, it is becoming common practice for old age psychiatrists to aim for lower lithium levels. However, this approach may not be known among general practitioners, who would naturally follow local laboratory advice and national NICE guidelines, which do not specify that lower plasma doses should be used in older people. Current British National Formulary (BNF) guidance (2025) states that lithium blood tests should be done weekly until concentrations are stable, then 3 monthly for the first year and then 6 monthly. ¹² The ISBD’s guidelines (2019), based on their Delphi survey, differ slightly in this regard and more pragmatically recommend 3–6 monthly monitoring of lithium. ^{Reference Shulman, Almeida, Herrmann, Schaffer, Strejilevich and Paternoster4}

If renal impairment develops, nephrology consultation is recommended, and the pros and cons of continuing lithium treatment must be weighed carefully.

On the basis of the somewhat limited literature on the topic, the results of published Delphi expert consensus studies and our own survey reported here, we consider that more conservative and safer lithium monitoring practices should be adopted in the UK for older people, in line with current clinical opinion, such as those suggested by the ISBD. ^{Reference Shulman, Almeida, Herrmann, Schaffer, Strejilevich and Paternoster4} Interestingly, in line with our views, the BNF recently changed its advice (in March 2025) and now suggests that initial serum lithium concentrations should be between 0.4 and 0.6 mmol/L for patients aged 65 years and older, adjusted according to response on specialist advice. This lower range is also specifically recommended in the prophylaxis of recurrent unipolar depression (for all ages). ¹² We accept that sometimes slightly higher levels may be needed in older people during a manic episode, so prefer the advice from the ISBD, but agree with these new initial levels recommended by the BNF.