Lithium-Associated Hypercalcemia: Pathophysiology, Prevalence, Management

Lithium has been used for a wide variety of ailments since the nineteenth century [1]. However, it was not until the illustrative case presentations of Cade [2] in 1949, and more convincingly with the work of Schou [3] in what may have been the first randomised study in psychiatry, that lithium became a more widely accepted medication in the treatment of mania in patients with bipolar disorder (BPD), previously known as manic depression. Organic disturbances associated with lithium salts were first documented by Garfinkel et al. [4] in 1973 who presented cases of both hypothyroidism and hyperparathyroidism in patients with concurrent lithium treatment. Since then, a series of articles have confirmed these initial observations. Studies concerning lithium and hypercalcemia are overwhelmingly retrospective in design and are limited in number (PubMed search with the words lithium AND hypercalcemia in January 2017 revealed only 139 articles). Lithium-associated hypercalcemia (LAH) is generally recognised as a prevailing side effect of lithium so long as treatment is necessary [5]. Lithium remains the golden standard for BPD, proving efficacious in the treatment of mania, as supplementary treatment for resistant depression and even reported as having anti-suicidal effects [6, 7]. Lithium is frequently prescribed as a chronic medication and is in most cases lifelong thus necessitating the regular monitoring of calcium homeostasis; clinically, this has not been done systematically, where the focus of regular blood work-ups has been lithium concentration, thyroid and kidney function. It has been suggested that individuals with BPD may have genetically modified calcium receptors, potentially leading to altered calcium levels [8]. Additionally, there exists controversy concerning the histopathological background and appropriate surgical strategy of lithium-treated patients who have developed biochemical signs of hyperparathyroidism [9, 10].

Data available for the assessment of hypercalcemia included total and corrected total calcium, creatinine and, where possible, parathormone. Parathormone was only available for bipolar patients with lithium (Table 1). Vitamin-D status was available in so few that comparison was not viable, and therefore abandoned. In addition, current medication was noted and classified according to whether the substance was a mood-stabiliser, central stimulants, antidepressant, anxiolytic, antipsychotic, anxiolytic other than benzodiazepines, and “other” medications most often pertaining to somatic disorders such as hypertension. Hypercalcemia was defined as average total calcium, TCa ≥ 2.50 mmol/l according to national guidelines.

Descriptive statistics included frequencies (percentages) and median values (ranges). Individuals with hypercalcemia were compared with those without hypercalcemia using Chi-square test (for categorical measures), and a nonparametric test on the equality of medians (for non-normally distributed continuous measures). The same tests were used to compare the distribution of the study covariates by the study groups (BP with lithium, BP without lithium and the control population). Unadjusted and adjusted logistic regression models compared odds of hypercalcemia in bipolar patients with ongoing lithium treatment as well as in a randomly selected control population with the corresponding odds in bipolar patients without lithium treatment.

The approval for the collection of medical data and the completion of this study was given by the Central Ethics Review Board at the University of Uppsala, Sweden (Dnr 2011/428).

Primary hyperparathyroidism is the third most common endocrine condition with a prevalence of approximately 0.5% in the general population, rising to 2–3% in postmenopausal women [12]. LAH is characterised by mildly elevated calcium levels (Fig. 2), which occur chronically or intermittently [13, 14]. Despite numerous articles highlighting the credible association between lithium and the development of hypercalcemia, the monitoring of parathyroid function has only recently been included in the revised recommendations of NICE [15] and the International Society for Bipolar Disorder [16]. In the light of this paucity of clear national and international guidelines, monitoring of parathyroid function has in many cases been scarce and irregular. This has led to evident difficulties in determining a reliable incidence and prevalence of LAH. McKnight et al. [17] reveal the absolute risk of developing LHPT is 10%, considerably higher than pHPT. Estimations of LAH or LHPT prevalence vary widely from 4.3 to 80%, though this possibly reflects the dimensions of study populations [18‐21]. The present study confirms previous observations that at least a quarter of lithium-treated patients have repeated disturbances in calcium homeostasis (Table 1).

Lithium is a mood stabilizer but the exact biochemical mechanism of action has yet to be fully elucidated. The lithium ion is thought to be active, particularly in signal transduction pathways at multiple sites. Two main pathways of mood-stabilizing action are proposed, namely the inhibition of inositol monophosphatase, IMPase (important for cell regulatory functions such as cell growth and apoptosis), and the inhibition glycogen synthase kinase, GSK-3 (important for apoptosis) [22‐24]. By inhibiting GSK-3, long-term plasticity and neurocellular protection and stability may be achieved [25].

These mechanisms may be involved in lithium’s implicatory role in the development of hypercalcemia. Szalat et al. [26] explains lithium’s involvement in the cascade leading to the inhibition of IMPase in parathyroid chief cells, resulting in intracellular changes in calcium levels and thereby parathyroid hormone (PTH) secretion. A widely accepted theory is that lithium interacts with the calcium-sensing receptor (CaSR), a ubiquitously expressed transmembrane G-protein-coupled cell surface receptor, which in turn leads to a so-called right-shift or set-point elevation of calcium concentrations in relation to PTH [27, 28]. Chief cells then react as if the extracellular concentration of calcium has decreased. Exactly how lithium interacts with CaSR is difficult to determine. Furthermore, increased direct secretion of PTH is thought to occur through the development of adenoma or hyperplasia; hypocalciuria likely arises through the inhibition of renal cyclic adenosine monophosphate (cAMP) [13, 29].

Lithium affects PTH secretion and thereby causes PTH-dependent hypercalcemia, whereas the use of drugs such as thiazides (not uncommon in the elderly) and vitamin D can cause PTH-independent hypercalcemia.

The pathoanatomical diagnosis of LAH is a matter of some contention. Primary HPT is caused by a single adenoma in 85% of cases, with hyperplasia in 10–15% [30]. Lithium, through GSK-3 inhibition, contributes to irregular Wnt/β-catenin signalling, believed to be significant for the development of parathyroid adenomas and hyperplasia [31]. Theoretically, lithium should exert an equal effect on all parathyroid glands [32]. Carchman et al. [9], however, state that there is no significantly increased risk for multiglandular disease (MGD) in lithium-treated patients. On the other hand, numerous studies present higher frequencies of MGD ranging from 25 to 83% [10, 26, 33‐37]. The surgical results in the present study further confirm these latter results with MGD occurring in five of seven patients (70%) (Table 4). This has clear ramifications for the treatment and management of LAH.

Four main strategies for management of LAH are available. Firstly, lithium may be discontinued, with calcium normalisation occurring in many though far from all cases [5]. Discontinuation may be difficult due to the serious risk of psychiatric deterioration. Secondly, careful surveillance may be an option, though the patient’s clinical status must be observed vigilantly. Thirdly, calcimimetic therapy has been described as possible strategy for LHPT, especially if continued lithium treatment is necessary [38]. This strategy is even proposed by Szalat et al. [26], particularly for those not suitable for surgery. The fourth and final strategy is surgery, though debate exists as to the degree of radicality that should be executed [9, 10, 33‐37].

Surgical management of LAH is influenced by whether the condition is considered to be primarily a single glandular disease (SGD) or multiglandular (MGD). Hitherto, studies give differing views as to whether bilateral neck exploration (BNE) or focused neck exploration (FNE) is the appropriate surgical strategy. The role of pre-operative localisation techniques, has not been conclusively established for patients with LHPT [39]. Lithium is considered an exacerbating factor that “de-masks” the predisposition for the development of an adenoma, and FNE is recommended [9, 40, 41]. Awad et al. [40], in a study population of 15, found only one patient with MGD. The authors of this study (ADM, GW, JJ), in material yet unpublished, have found that four (1.5%) of 297 patients had elevated calcium levels (> 2.5 mmol/l) before the initiation of lithium treatment. The possibility of two coexisting phenomena, i.e., lithium treatment and pHPT, without any causal relationship, has to be considered [42]. Nevertheless, most studies suggest the frequent occurrence of MGD [10, 26, 34‐36]. Hundley et al. [10] further point out that the resection of a single gland increases the risk of disease recurrence. On the other hand, the risk of permanent hypoparathyroidism with more radical surgery, though low, is not negligible [37]. The present recommendations of the European Society of Endocrine Surgeons (ESES) suggest that a history of lithium treatment is a relative indication for BNE [29].

There currently exists a lack of systemization in the monitoring of parathyroid function in lithium-treated patients, which may reflect the lack of clear guidelines on both the national and international level. Regular monitoring, involving TCa, PTH and vitamin D, would potentially identify all patients with tendencies towards hypercalcemia, whereby appropriate action could be taken. Hypercalcemia in lithium-treated patients with BPD was not correlated with kidney function, and even moderate chronic kidney insufficiency proved to be statistically insignificant (Table 3). The significance of hypercalcemia for the lithium-treated individual is unclear [14]. The GAF instrument did indicate that individuals with lithium treatment scored lower on this assessment of psychosocial functioning (Tables 1, 2). It illustrates the difficulty in discerning and differentiating symptoms primarily related to bipolar disorder and those associated with hypercalcemia. No robust instruments exist to differentiate these two conditions, but the need is apparent. It is often thought that non-lithium-treated patients with mildly elevated calcium are “asymptomatic”, often leading to “doctor’s delay” [19]. Talpos et al. [43], using SF-36, reported dramatic improvements in quality of life for seemingly asymptomatic HPT patients undergoing parathyroidectomy, effects that remained 2 years post-operatively. This underlines the need to penetrate the patient’s anamnesis in a structured way in that it may play an important role in deciphering indications for surgery for LAH patients.

A large percentage of lithium-treated patients may be at risk of developing hypercalcemia, but far from all [31]. Some patients can medicate with lithium for decades without, at any time, showing tendencies towards calcium elevation. The reason for this is unclear. In the present study, lithium treatment, age, gender and thyroid disease were strongly associated with the development of hypercalcemia, while studies even point to treatment duration and dosage as associated risk factors [5, 14, 18, 28]. These factors should be considered in any assessment for surgery as they may prove important in deciding the extent of any eventual operation.

The question of surgical management of LAH needs to be explored further. Most cases presented here had MGD (Table 4), though on average only two parathyroid glands were removed. The extent of surgery at the initial operation may have importance for the risk of recurrence, particularly for patients continuing lithium medication. There exists a strong need for prospective randomized surgical studies to evaluate appropriate management strategies in relation to patient benefit. Prospective therapeutic studies are also needed, including the evaluation of calcimimetics.

This study has several limitations. It is cross-sectional and thereby bereft of information as to disease debut or how calcium homeostasis changes over time. Secondly, it is retrospective, which adds to the risk that the collected information is selective and incomplete. Thirdly, though we recommend monitoring PTH and vitamin D, we have not been able to collect PTH for all the subgroups and vitamin D for only a small number of individuals. Collection of this data would have strengthened any conclusions that are made.