Decline in prevalence of lithium prescriptions in older adults

Little is known about patterns of lithium prescribing to older adults. To illustrate age-specific prescribing trends, we analysed routinely collected UK primary care data from the Clinical Practice Research Datalink (CPRD) Gold and Aurum databases, which together cover 15–20% of the UK population and are broadly representative in terms of age, sex, ethnicity and deprivation. ^{Reference Sanchez-Santos, Axson, Dedman and Delmestri5} Adults aged 65–100 years with any record of a bipolar disorder episode or diagnostic code between 2005 and 2022 were identified using validated diagnostic codes. ^{Reference Launders, Richards-Belle, Hardoon, Osborn and Hayes6} A total of 25 954 patients were included in the cohort. Of these, 2441 patients (9.4%) had at least 1 lithium prescription during follow-up. The annual prevalence of lithium prescribing fell steadily from 13.8% in 2005 to 6.3% in 2022, roughly halving over this period (Fig. 1).

Fig. 1

Annual prevalence of lithium prescribing among adults aged 65–100 years with a recorded bipolar disorder diagnosis in the UK from 2005 to 2022. Annual prevalence of lithium prescribing was calculated for each calendar year as the number of patients with at least one lithium prescription that year (numerator), divided by the number of patients with at least one day of follow-up in that year (denominator), expressed as a percentage. Follow-up began on 1 January 2005, the date of registration plus 365 days, the year the patient turned 65 years old or the date of the first bipolar disorder code recorded within the study period (whichever was latest); follow-up ended on 31 December 2022, the year the patient turned 100 years old, or upon Clinical Practice Research Datalink deregistration, practice last collection date or death (whichever was earliest).

The evidence gap

Lithium remains the gold-standard treatment for bipolar affective disorder. However, there is a paucity of studies in OABD, in part because older adults are underrepresented in randomised controlled trials (RCTs) and often excluded from trials because of comorbidity, frailty or concomitant medication use. The 2017 GERI-BD (Acute Pharmacotherapy in Late-Life Mania) study represents the only RCT conducted in older people with bipolar disorder. ^{Reference Young, Mulsant, Sajatovic, Gildengers, Gyulai and Al Jurdi7} This double-blind RCT compared the efficacy and tolerability of lithium and valproic acid in 224 individuals aged over 60 years with mania or hypomania. Lithium demonstrated superior efficacy in reducing mania symptoms at 9 weeks, and the two drugs were similarly well tolerated. However, GERI-BD was a trial in individuals with acute mania with a follow-up of only 9 weeks. ^{Reference Soh, Drori and Rej2} Dedicated RCTs evaluating maintenance treatment in OABD are lacking, and current recommendations are inferred from RCTs conducted in younger adults. ⁴

The risk gap

Lithium has a narrow therapeutic window in terms of serum concentration. However, there remains limited evidence to determine recommendations regarding the optimal age-specific therapeutic range in older adults. An expert consensus from the International Society for Bipolar Disorder Older Adults Task Force recommended maintenance serum lithium levels of 0.4–0.8 mmol/L for patients aged 60–79 years and 0.4–0.7 mmol/L for those aged 80 years and over. ^{Reference Rubinsztein, Willis, Birgi, Sachdev, Southwell and Stewart1} An International Society for Bipolar Disorder Task Force systematic review of serum lithium concentrations in OABD reported mean or median lithium concentrations ranging from 0.47 to 0.67 mmol/L. ^{Reference Nunes, Olagunju, Zittema, Dhyani, Fu and Chen8} Similarly, a survey study conducted by the Royal College of Psychiatrists’ Faculty of Old Age Psychiatry recommended an initial target serum lithium concentration of between 0.4 and 0.6 mmol/L for adults aged 65 years and older. ^{Reference Rubinsztein, Willis, Birgi, Sachdev, Southwell and Stewart1} However, these recommendations were derived from observational data, a Delphi study and survey data rather than higher-quality forms of evidence.

Elevated lithium levels are associated with acute toxicity, which may present with neurological symptoms including delirium and cerebellar dysfunction, as well as gastrointestinal disturbance. ^{Reference Nunes, Olagunju, Zittema, Dhyani, Fu and Chen8} Other common adverse effects include polydipsia, polyuria and weight gain, whereas longer-term treatment may be associated with hypothyroidism and hypoparathyroidism. Older adults are considered to be at greater risk of lithium toxicity, but there is a lack of data showing direct age comparisons. Older patients are also more likely to experience polypharmacy, increasing the risk of drug interactions with lithium. These potential adverse effects may disproportionately influence prescribing decisions in older adults, despite the fact that many can be monitored and managed safely within routine clinical practice.

The impact of long-term lithium exposure on renal function is also a common concern when prescribing lithium in older adults. In a Swedish cohort study, long-term lithium exposure was found to be associated with a steeper decline in renal function. However, the magnitude of this effect was modest: the estimated glomerular filtration rate declined by 0.57 mL/min/1.73 m² per year in the general population, compared with 0.99 mL/min/1.73 m² per year among individuals exposed to lithium for more than 10 years. ^{Reference Fransson, Werneke, Harju, Öhlund, de Man Lapidoth and Jonsson9} Further, there was considerable inter-individual variation, with some individuals maintaining preserved renal function after several decades of treatment. Comorbidities such as hypertension and diabetes also contributed to kidney disease. Regular monitoring of renal function and serum lithium concentrations may help to mitigate this risk and facilitate earlier identification of individuals with accelerated renal decline. ^{Reference Soh, Drori and Rej2} Additional strategies include once-daily lithium dosing and pro-active management of comorbid conditions such as hypertension and diabetes. ^{Reference Shivakumar, Szymaniak, Bell, Boyce, Porter and Malhi3}

In older adults, in whom a decline in renal function with age may be further affected by comorbidities, determining when lithium should be discontinued is a challenge. Notably, there is no universally accepted threshold for when discontinuation should occur. The National Institute for Health and Care Excellence (NICE) advises consultation with a nephrologist when the baseline estimated glomerular filtration rate is <60 mL/min/1.73 m² or when there is a clinically significant sustained decline in renal function. ⁴

The system gap

System-level barriers to lithium prescribing are likely to vary among healthcare systems. The healthcare pathway must be able to facilitate physical health investigations required before commencing lithium treatment and ongoing monitoring requirements. Monitoring requirements may be more challenging in older adults, for whom more frequent surveillance of serum levels is recommended (typically every 3 months compared with every 6 months in younger adults). ⁴

In the UK, responsibility for lithium prescribing and monitoring is frequently shared between primary and secondary care, and there may be uncertainty regarding which clinician or service assumes overall responsibility. These shared-care agreements also vary by geographical region and local commissioning arrangements, creating additional complexity around ownership of monitoring and prescribing decisions. Consequently, clinicians may be reluctant to initiate lithium unless they are confident that robust systems are in place to support safe long-term follow-up.

Although NICE recommends using lower lithium doses in older adults, the absence of lithium carbonate 50 mg and 100 mg tablet formulations in the UK can make this difficult to achieve in practice. ⁴ Whereas lithium citrate liquid may provide an alternative means of achieving lower doses, liquid formulations may be less practical for some older adults, particularly those with visual, cognitive or manual dexterity difficulties. Whether this represents a wider international prescribing barrier is unclear, as lithium formulations vary among countries.

Given the practical challenges in initiating and maintaining lithium safely, it is perhaps not surprising that clinicians have increasingly turned towards alternative treatments such as second-generation antipsychotics or alternative mood stabilisers. ^{Reference Shivakumar, Szymaniak, Bell, Boyce, Porter and Malhi3} These do not require the same intensity of monitoring. Newer medications may also be more prominently marketed and advertised. The majority of acute mania episodes are treated with antipsychotics. Furthermore, most acute mania episodes are now treated with second-generation antipsychotics rather than lithium. This may present an additional barrier to subsequent lithium use, as patients who respond well to an antipsychotic during the acute phase may be more likely to continue the same treatment into the maintenance phase rather than switching to lithium.