Lithium was discovered as a therapeutic remedy for psychiatric conditions in the mid-19th century and was reintroduced one century later [1] and it is still the most widely used medication for long-term management of bipolar disorder, where it is administered as a salt in the form of lithium carbonate/cirate/chloride/or sulfate. Bipolar disorder (BD) is a serious life-long disorder, characterized by recurrent episodes of depression and mania [2]. BD is classified, based on the presence of depressive along with manic or hypomanic episodes, into bipolar type I and type II disorder [3]. In bipolar II disorder, depression predominates, and the manic episodes are milder and briefer. This milder and less prolonged form of mania is referred to as hypomania. A person with bipolar I disorder, however, will experience a full manic episode and may or may not experience a major depressive episode [3]. BD affects 2.4% of the world population and is a leading cause of disability worldwide [4]. In its more severe forms, BD is associated with significant impairment of personal and social functioning and high risk of death through suicide as well as poor physical health. Lithium, having the strongest evidence of long-term relapse prevention, is the first-line treatment for both acute and maintenance treatment of BD. Furthermore, lithium is also prescribed for major depressive disorder as an adjunct therapy, as well as a treatment of vascular headaches and neutropenia [5,6]. In addition to its mood-stabilizing properties, remarkable neuroprotective and antiviral properties have also been attributed to lithium, with the use of lithium recently proposed as a potential treatment for Covid-19 [7]. Altogether, it is estimated that up to one million people worldwide take lithium on a daily basis [8].

Despite its global therapeutic use, the benefits of lithium are restricted by its narrow therapeutic index and the incidence of adverse effects [9]. The narrow margin between the safe and potentially toxic doses of lithium has resulted in self-administration of toxic doses accounting for 20–27% of hospitalized poisoning [9], and mortality rates of 9 to 25% reported from lithium toxicity [10]. The therapeutic effect of lithium salts is directly associated with its level in blood serum and there exist differences in individual pharmacokinetic and risk of intoxication [11]. Based on the differences in the excretion rates between individuals, the daily lithium dosage can vary between 10 and 80 mmol, which results in plasma concentrations of 0.4–1.2 mmol/L for effective treatment [12]. Lithium plasma levels greater than 1.2 mmol/L are potentially toxic and can be fatal. Therefore, avoidance of lithium intoxication has been, and continues to be, an important component in lithium treatment, and lithium serum levels must be monitored constantly to ensure its effectiveness and prevent adverse effects [9]. Lithium toxicity is associated with neurotoxic events, hyperthyroidism, hypercalcemia, and other serious conditions [10]. Considering the potential consequences of lithium toxicity, vigilant monitoring should be central in the treatment of BD patients. In the majority of cases, lithium toxicity is preventable with regular monitoring which can significantly reduce the number of toxic episodes in lithium-treated patients [9].

Additionally, treatment non-adherence is a persistent problem in psychiatry, with about 54% of patients not adhering to their prescription [10]. Despite lithium’s proven benefits regarding the prevention of severe affective episodes and suicide, discontinuation of lithium treatment is common amongst bipolar patients with about half of all individuals on lithium medication stopping their treatment at some point, which results in high levels of relapse. Along with psychiatric and physical reasons interfering with lithium treatment, lithium discontinuation has been suggested to be mainly due to its adverse effects which substantially impair the quality of life. Common adverse effects leading to lithium discontinuation are diarrhea, tremor, polyuria/polydipsia/diabetes insipidus, creatinine increase, and weight gain [13]. Therefore, side effects, toxicity burden associated with lithium medication, and the need for regular monitoring via vein puncture are the main reasons for lithium discontinuation and treatment non-adherence [9]. Consequently, the toxicity and treatment non-adherence burden associated with lithium medication has resulted in the widening of the mortality gap between BD patients and the general population. It is hence crucial to develop strategies to improve adherence and prevent unnecessary termination of lithium treatment [13].

Reaching and sustaining the right therapeutic level to avoid toxicity, dose-related adverse effects, and consequently, treatment non-adherence, requires regular therapeutic monitoring of lithium concentrations. In general, peak lithium concentrations in plasma occur two to four hours after an oral dose, with complete absorption occurring at around eight hours. Therefore, it is crucial to monitor lithium levels in serum twelve hours after the last dose [12]. Currently, lithium determination is performed by withdrawing blood samples from the patient by trained personnel, which is invasive and often painful, especially in patients with difficult venous access. The sample is then transported to the central laboratory where blood cells are removed before measurement, a procedure that can take up to 45 min. The quantification of lithium in clinical laboratories is performed by atomic absorption spectrometry (AAS), flame emission photometry (FEP), or conventional ion-selective electrodes (ISEs) [9,14]. These techniques are costly and require highly qualified personnel and elaborate laboratory methods that cannot be translated into point-of-care devices for personal monitoring. Moreover, with the current techniques, the whole analytical process of adjusting the dose after the first administration may have a variable lag time, usually from a few days to weeks [15]. The delay between sample extraction and the analysis of the results restricts the possibility of early detection, correction of problems, and prevention of potential adverse effects.

Employing optical techniques for monitoring is vital for BD patients to have comprehensive informlithium levels offers numerous advantages including high accuracy, specificity, and good biocompatibility. Nevertheless, development of optical lithium sensors is associated with some challenges. For example, the use of optical lithium ligands is limited by the intrinsic properties of lithium cation such as its small size and its high charge density which can result in poor coordination about their pharmacologicalility. Therefore, the right optical lithium ligand should be identified to allow selective and efficient sensing of lithium [55,66]. Furtreatment at the appropriate time as it can substantially imprhermore, in the development of optical lithium ligands, the response time for signal detection should be considered. The response time of small ligands used for lithium detection has been reported to vary from a few seconds to 30 min, with most of the optical sensors displaying detection after 1–10 min, because of the time needed for Li⁺ catiovn to coordinate treatment adwith ionophore, and the necessary duration for stabilizing the detection signal [50]. Therefore, the equilibrationce [13]. time for Li⁺ Pcoordinatients’ percepon must be considered which depends on the ligand that is employed in the membrane for preparation of side effects or apprehensthe optical sensor. The reversibility of ligands for lithium detection is another crucial parameter to be considered. This is because after Li⁺ detection ofthe lithium intoxication, as opposed to the actual pressensor is classically non-reusable and should only be reused in the condition of a total de-coordination of lithium cation from coordination sites and the recovery of initial optical properties [50]. Ultimatencly, the of side effects, may furptical lithium detection/quantification must not or must only slightly affect the studied biological system. Lithium der contritermination using FES, FAAS, and ISE methods has been evaluated [68]. It has been sute to treatment non-adherenceggested that although the percentage of lithium recovery for FES and FAAS methods was high, ISE showed the most satisfactory results for recovery of lithium in pooled patient’s serum [968]. TNevertheless, it is suggested that FES remains therefore, there is a crucial need to have quicker and simpler ways to preferred method for lithium measurement in many laboratories. This is mainly due to the lack of standardization and validation of novel technologies which calls for a strong collaboration between academic, industrial, and medical partners. Moreover, it is demonstrated that a higher average lithium concentration for patients’ serum samples was measured by ISE and that ISE determine lithium levelsation is easier and more precise. Nonetheless, the accuracy of ISEs may depend on other interfering factors [68]. DFor evelopment of a minimally invasive xample, one of the main challenges associated with ISEs for lithium- monitoring method will bis the ongoing research on finding suitable ionophores with sufficiently high selectivity in respect to sodium [41,42]. In general, a major advance ccurate measurement of lithium concentration by ISE requires either very high selectivity of the membrane for Li⁺ or an alithium gorithm that will compensate for interference from sodium [41]. Furthermore, the electrochemicapeutic monitoring as itl (EC) sensors giving a current response directly proportional to analyte concentration are preferred to potentiometric ISEs, which give a potential response that is proportional to the logarithm of analyte activity [80]. In generallows patients and non-medically-trained personnel to measure, ISEs have been mainly coupled with potentiometric and voltammetric detection methods which both suffer from limitations such as drifting of electrode-electrolyte interface impedance. Employment of electrical impedance spectroscopy is also an interesting but understudied approach for lithium levels and ensure patients are receiving the optimum dose [16]. Moreover, monitoring of drugs with narrow therapeutic range at home using wearable detection. While tetrapolar electrical impedance spectroscopy offers a high degree of sensitivity to conductivity variations in samples, it lacks sensitivity towards lithium ion and must be combined with other techniques such as optical methods. Tripolar configurations, measuring the properties of functionalized electrodes, can also be investigated as they might offer a higher degree of selectivity for lithium determination. Solid-contact ion-selective electrodes (SC-ISEs) also have some limitations, including the need for calibration, limited selectivity, poor reliability, and potential drift [81,82]. Lastly, mensors can help reduce the burden on patients and heasuring compounds using capillary electrophoresis (CE) can also be challenging due to wall adsorption shifting performance, particularly over multiple runs. Therefore, different types of surface coatings have been investigated to reduce issues relating to adsorption effects [42]. Capillth professionary electrophoresis microchips [37,77,79] also associated with attenhave several disadvantages, including complicated channel design and fabrication procedures and high-voltage manipulation using expensive instruments [46]. Moreover, band-broance of clinical settings and facildening in CE is a general problem for samples containing a high concentration of ionic constituents as it can lead to poor resolution [77]. An importate improvednt step in the development of a point-of-care (POC) device for therapeutic monitoring [17]. Improved rates of lithium of lithium is matrix selection. Direct analysis of whole blood without any sample pretreatment moniremains one of the main challenges in blood monitoring will majorly enhanc. Therefore, combining sample treatment steps with the sensing methodologies on a single device can facilitate the quality of life of bipdevelopment of a point-of-care device for lithium monitoring [42]. Blood-relar patients and will allow bipolarted matrices have some major limitations; namely, blood sampling is relatively invasive and may be impractical for certain patients to monitor the drug during th populations. The costs associated with collecting, transporting, and processing blood samples are also significant [42]. As aforementir treatment rathoned, dried blood spot (DBS) and dried plasma spot (DPS) have been investigated for lithium analysis [21]. However, than taking single measurements weekly or monthly, which will greatly reduce the risk of adverse effects during the course of their treatment. The pivotal need for the development of low-cost and easy-to-uswhen using DBS and DPS to alleviate large-volume venous sampling, the influence of temperature, humidity, and sunlight exposure must be considered. Additionally, patients or caregivers must be well instructed to lower the chance of sample contamination. Several biological fluids such as sweat, saliva, and ISF have been investigated as alternative matrices to replace blood-based approaches. Nevertheless, the accumulated knowledge on drug concentration and therapeutic response dynamics needs to be established for different matrices to correlate drug levels with blood and plasma concentrations. Several studies have investigated saliva for lithium therapeutic monitoring. Although saliva can be collected easily even without patient stimulation, sensing in saliva is challenging as plaques and bacteria might attach on the sensors for decentralized monitor surface and compromise the efficacy of the system. Moreover, another limitation is the timing of lithium has resusaliva sampling since equilibrium of analyte transport between the blood and saliva must be present [83]. Lastly, ted in a great interest in he concentrations in saliva are strongly dependent on dietary intake. ISF, providing a medium between vasculature and cells and rich in molecules and metabolites targeted in TDM, has also been investigating noveled for lithium- monitoring techniques. Efforts have been made to develop minimally invasive point-of-care lithium-monitoring devices utilizing m. Nonetheless, one of the common challenges with ISF is the extraction of a reasonable amount of analytes for downstream analysis. Careful consideration must also be given to microneedles or under-the-skin excitation currents which are used to take out interstitial fluid, as these methods can cause irritation and discomfort in the dermis layer, especially during prolonged monitoring [31]. Sweat-based apprices such as blood, urine, sweat, salivaoaches for lithium TDM suffer from several issues as well. For example, the analyte concentration profiles may be location dependent as sweat glands are unevenly distributed, the sweat glands need to be stimulated by exercise or thermal heating to achieve analyte collection, and interstitial fluid (ISF). The current paper reviews the literature onthere can be high inter- and intra-individual variability. Other constraints regarding drug stability, concentration accuracy, sweat evaporation, and chance of contamination during the sampling process must also be considered. Finally, urine, which is investigated technologby several studies for non-invasivelithium monitoring of lithium medic, requires sensitive analytical methods and is limited by the wide variation in the treatment of BDbetween and within patients.