Sage Journals: Discover world-class research

Abstract

Objective:

Hypothyroidism is a well-documented consequence of lithium treatment. Less well known is a possible association between lithium therapy and hyperthyroidism. This may have clinical implications as rapid changes in thyroid hormones may worsen a person’s affective state, while symptoms of hyperthyroidism can mimic those of mania. We therefore systematically reviewed the published literature for evidence of lithium-induced hyperthyroidism.

Methods:

We searched PubMed, Embase and CINAHL for articles where individuals developed biochemically confirmed hyperthyroidism (with or without clinical symptoms), while on lithium therapy for an affective illness. We included case reports, case series, cross-sectional, case control and cohort studies.

Results:

We included 52 studies, 39 of which were individual case reports and 3 were case series. There were 10 cross-sectional or case control or cohort studies. All the research designs suggested an association between the prescription of lithium and hyperthyroidism. However, these findings were limited by the quality of the included studies, small number of participants and the general lack of either a clear temporal relationship or dose response.

Conclusion:

Hyperthyroidism is an uncommon side-effect of lithium compared to hypothyroidism but may have clinical implications. However, large prospective studies are required to clarify this association and to further inform the management of patients treated with lithium where hyperthyroidism occurs.

Keywords

Lithium hyperthyroidism thyrotoxicosis bipolar affective disorder schizoaffective disorder

Introduction

Hypothyroidism caused by lithium treatment is now well documented (Bocchetta et al., 2007b; Bocchetta and Loviselli, 2006; Bou Khalil and Richa, 2011; Kibirige et al., 2013). Less well known, and somewhat surprising given its antithyroid mechanism of action, is an association between lithium therapy and hyperthyroidism.

The first case report appeared in 1974, and since then, a growing body of reports has added to the evidence base (Franklin, 1974). Although there is a causative relationship between lithium therapy and other forms of thyroid illness (goitre and hypothyroidism; Lazarus, 2009; Lazarus et al., 1986), the nature of the association between lithium and hyperthyroidism is less clear. Possible mechanisms include a direct toxic or immuno-stimulatory effect on the thyroid (Kibirige et al., 2013).

One study identified thyrotoxicosis in 3.9% of women and 1.8% of men treated with lithium therapy, which was similar to rates of hyperthyroidism reported in a 20-year study of the incidence of thyroid disease in a representative sample of the United Kingdom (Kirov et al., 2005; Vanderpump et al., 1995). Incidence rates per 1000 survivor years on lithium were 6.4 for women and 2.7 for men (Kirov et al., 2005). A further study reported an incidence rate for lithium-associated thyrotoxicosis of 2.7 cases per 1000 person-years, while that for silent thyroiditis was approximately 1.3 cases per 1000 person-years (Miller and Daniels, 2001). The latter is a transient thyrotoxicosis with a very low 24-hour radioiodine uptake. Barclay et al. (1994) found that the number of cases of thyrotoxicosis in patients on lithium was more than three times that predicted from the local thyrotoxicosis incidence rate. However, these calculations were made based on an estimate of the number of patients in the area on lithium, which was derived by extrapolating from old data dating from 1974, and the amount of lithium sold in the area in 1989 (Barclay et al., 1994). More recently, a population-based cohort study found that the rate of hyperthyroidism among lithium-treated patients increased compared to olanzapine and valproate, but not quetiapine (Hayes et al., 2016).

By contrast, a retrospective analysis of laboratory data (Shine et al., 2015) and a cross-sectional comparison of those exposed to lithium therapy and controls did not find any association between that lithium therapy in general and hyperthyroidism (Kuman Tuncel et al., 2017). However, the former study did report an association in patients with higher than median lithium concentrations (Shine et al., 2015).

In summary, most case reports, case series and cohort studies suggest that exposure to lithium treatment is associated with an increased risk of hyperthyroidism. However, some authors do not report any association. Although uncommon, hyperthyroidism is an important and clinically relevant phenomenon as rapid changes in thyroid hormones may worsen a person’s affective state while symptoms can mimic those of mania (Arlt et al., 2008; de Sousa Gurgel et al., 2015; Iga et al., 2005; Szabadi, 1991). We therefore systematically reviewed the published literature for evidence of lithium-induced hyperthyroidism.

Methods

Search strategy

The systematic review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; Moher et al., 2009). We sought advice from a research librarian prior to the commencement of the search. Articles were located by searching PubMed (1964 to January 2018), Embase (1947 to January 2018) and CINAHL (1937 to January 2018). We restricted our search to English-language articles. Table 1 shows the search terms. The review protocol is available from the authors. Ethical approval was not required for this systematic review, as all included primary data had been previously published with ethical approval.

Table 1.

Search strategy.

Database	Search terms
PubMed (n = 199)	‘Hyperthyroidism’[Mesh] OR Hyperthyroidism[tiab] OR ‘thyrotoxicosis’[MeSH Terms] OR ‘thyrotoxicosis’[tiab] OR hyperthyroidism OR ‘silent thyroiditis’ OR thyroiditis ‘Adverse effects’[Subheading] OR harm[tiab] OR harms[tiab] OR side-effect[tiab] OR ‘side-effects’[tiab] OR Undesirable effect[tiab] OR undesirable effects[tiab] OR undesirable reaction[tiab] OR undesirable reactions[tiab] OR undesirable event[tiab] OR undesirable events[tiab] OR undesirable outcome[tiab] OR undesirable outcomes[tiab] OR adverse effect[tiab] OR adverse effects[tiab] OR adverse reaction[tiab] OR adverse reactions[tiab] OR adverse event[tiab] OR adverse events[tiab] OR adverse outcome[tiab] OR adverse outcomes[tiab] ‘Lithium’[Mesh] OR lithium OR lithicarb OR eskalith OR lithobid OR lithane OR cibalith-s OR quilonum OR hypnorm Filter: Human, English language
Embase (n = 138)	(‘hyperthyroidism’/exp OR Hyperthyroidism:ti,ab OR ‘thyrotoxicosis’/exp OR ‘thyrotoxicosis’:ti,ab OR hyperthyroidism OR ‘silent thyroiditis’ OR thyroiditis) AND (‘adverse effects’/exp OR harm:ti,ab OR harms:ti,ab OR side-effect:ti,ab OR ‘side-effects’:ti,ab OR ‘undesirable effect’:ti,ab OR ‘undesirable effects’:ti,ab OR ‘undesirable reaction’:ti,ab OR ‘undesirable reactions’:ti,ab OR ‘undesirable event’:ti,ab OR ‘undesirable events’:ti,ab OR ‘undesirable outcome’:ti,ab OR ‘undesirable outcomes’:ti,ab OR ‘adverse effect’:ti,ab OR ‘adverse effects’:ti,ab OR ‘adverse reaction’:ti,ab OR ‘adverse reactions’:ti,ab OR ‘adverse event’:ti,ab OR ‘adverse events’:ti,ab OR ‘adverse outcome’:ti,ab OR ‘adverse outcomes’:ti,ab) AND (‘lithium’/exp OR lithium OR lithicarb OR eskalith OR lithobid OR lithane OR cibalith-s OR quilonum OR hypnorm)
CINAHL (n = 12)	(MH‘Hyperthyroidism’ OR Hyperthyroidism OR MH ‘Thyrotoxicosis’ OR ‘thyrotoxicosis’ OR hyperthyroidism OR ‘silent thyroiditis’ OR thyroiditis) AND (MH‘Adverse Drug Event’ OR harm OR harms OR side-effect OR ‘side-effects’ OR ‘undesirable effect’ OR ‘undesirable effects’ OR ‘undesirable reaction’ OR ‘undesirable reactions’ OR ‘undesirable event’ OR ‘undesirable events’ OR ‘undesirable outcome’ OR ‘undesirable outcomes’ OR ‘adverse effect’ OR ‘adverse effects’ OR ‘adverse reaction’ OR ‘adverse reactions’ OR ‘adverse event’ OR ‘adverse events’ OR ‘adverse outcome’ OR ‘adverse outcomes’) AND (MH‘Lithium’ OR lithium OR lithicarb OR eskalith OR lithobid OR lithane OR cibalith-s OR quilonum OR hypnorm)

Two of the authors (F.B. and N.P.) independently screened identified records and abstracts. The reference lists of all included studies were screened to identify additional relevant studies that met inclusion criteria.

Inclusion criteria

We included articles where individuals developed biochemically confirmed subclinical or clinical hyperthyroidism while on lithium therapy for an affective illness (e.g. bipolar disorder, schizoaffective disorder and depression). Where older psychiatric diagnostic terms were used, such as ‘manic psychosis’, these were converted into a standardised diagnosis congruent with the International Classification of Diseases – 10th Revision (ICD-10) or the Diagnostic and Statistical Manual of Mental Disorders (5th Edition; DSM-5), based on the reported clinical history. Patients were included if they had an existing thyroid illness such as a goitre or hypothyroidism, unless they were hyperthyroid prior to commencing lithium therapy. The decision to include cases of premorbid thyroid illness was made on the premise that lithium use could accelerate the patient’s transition to hyperthyroidism leading to illness in predisposed individuals (Chow et al., 1993). We included the following study designs: case series, cohort studies, case control and cross-sectional studies.

Exclusion criteria

We excluded papers where individuals developed hyperthyroidism after the withdrawal or cessation of lithium therapy. This is because these cases may have had a pre-existing hyperthyroid illness that was masked by the thyroid-suppressing effects of lithium, and which then became apparent on cessation of this treatment (Carmaciu et al., 2003; Miller and Daniels, 2001).

Study quality

We assessed the quality of cross-sectional, case control or cohort studies using a modified version of the Newcastle–Ottawa Scale (Mata et al., 2015). This scale assesses the quality of non-randomised studies included in systematic reviews and meta-analyses across the following domains: sample representativeness and size, comparability between respondents and non-respondents, ascertainment of hyperthyroidism and statistical quality (Mata et al., 2015). We divided studies into those at low risk of bias (⩾3 points) or high risk of bias (<3 points).

Statistical analysis

The limited number of studies and their variations in methodology, particularly length of follow-up, meant that statistical pooling in a meta-analysis was neither possible nor appropriate. Results were therefore restricted to a narrative synthesis. We extracted descriptive statistics for each type of study design including the age, sex, psychiatric diagnosis, duration of lithium treatment prior to thyrotoxicosis, lithium dose and level at the time of presentation with thyrotoxicosis, and evidence of the following: (1) clinical signs/symptoms of hyperthyroidism, (2) pre-existing thyroid illness and (3) presence of antithyroid antibodies. Medical comorbidities were not included as they were seldom recorded.

Categorisation of study types

Cohort studies were defined as observational studies, where the sampling of study participants was based on exposure to lithium. The starting point for a cohort study may be all exposed or all unexposed patients, or both. Notably, a comparison group is not a defining feature of a cohort study (Dekkers et al., 2012). By contrast, case studies/series case control studies were observational designs, which consisted solely of participants who were sampled based on the presence of a disease or disease-related outcome (Dekkers et al., 2012).

Results

We identified 349 studies in the databases (199 in PubMed, 138 in Embase and 12 in CINAHL), with a further 11 studies found via screening of references or other sources (Figure 1). After assessing papers at the title and abstract level, we included 52 studies, 39 of which were individual case reports and 3 case series (total n = 42). There were 10 cross-sectional, case control or cohort studies (Figure 1).

Figure 1.

PRISMA flow chart of included and excluded studies.

The 39 case reports gave details on 45 patients (Table 2) (Altieri et al., 2015; Arlt et al., 2008; Bafaqeeh and Myers, 1976; Bandyopadhyay and Nielsen, 2012; Becerra-Fernandez, 1995; Bernstein and Friedman, 2011; Byrne and Delaney, 1993; Chalasani and Benson, 2014; Chow et al., 1993; Cubitt, 1976; Dalan et al., 2007; Dang and Hershman, 2002; De Sousa Gurgel et al., 2015; Deardorff et al., 2016; Dwarakanathan, 1998; El-Bakush et al., 2014; Fauerholdt and Vendsborg, 1981; Franklin, 1974; Humphreys and Waddell, 1988; Kar et al., 2014; Law et al., 2004; McDermott et al., 1986; MacGregor, 1977; Merry, 1977; Mizukami et al., 1995; Numata et al., 2002; Onuigbo et al., 2000; Pallisgaard and Frederiksen, 1978; Rabin and Evans, 1981; Reus et al., 1979; Sato et al., 2013; Shimzu et al., 1997; Sinnott et al., 1992; Siyam et al., 2013; Tan et al., 2013; Todd and Jerram, 1978; Valenta et al., 1981; Yamagishi and Yokoyama-ohta, 1999; Yassa et al., 1988).

Table 2.

Case reports.

No.	Author(s), year	Age (years)/ sex	Diagnosis	Lithium dose	Li level	Symptoms	Auto-antibodies	TSH	T3	T4	Sequence thyroid dysfunction	Treatment
1	Franklin (1974)	45/F	Depression, 20 years	4 year Li therapy, pre-existing non-toxic goitre – euthyroid 1 g	Not reported	Tremor, sweating, heat intolerance, weight loss, firm goitre	Not reported	Not reported	Not reported	Not reported	Histology toxic goitre	Thyroidectomy
2	Cubitt (1976)	59/F	Cyclothymic PD	Not reported	Not reported	Tremor, no goitre	Not reported	Not reported	Not reported	Total = 19.5–24.4 umol/L (5.6–14.3), Free T4 index = 0.170–0.231 (0.049–0.138)	Thyrotoxicosis	Carbimazole
3	Bafaqeeh and Myers (1976)	50/M	BPAD	3 year duration, 0.5–1.0 g	0.6 mmol/L	Agitated, tremulous, sweating, tachycardia, perceptual disturbance and paranoia	Not reported	Not reported	210 nmol/L (55–150)	Not reported	Thyrotoxicosis	Not reported
4	MacGregor (1977)	39/F	Depression	5 years	Not reported	Weight loss, tremor, lid lag, no goitre	Negative	Not reported	Not reported	T4 210 mmol/L	Hyperthyroidism	Carbimazole
5	Merry (1977)	54/M	Depression	1200 mg (2 years)	Not reported	Shakiness, dizziness, palpitations, weight loss	Not reported	Not reported	1.47 nmol (0.8–1.5)	257 nmol (50 to 1–154 nmol)	Thyrotoxicosis	Carbimazole
6	Todd and Jerram (1978)	(a) 36/F	Depression	800 mg, 0.58	0.58, second level 0.36	Weight loss, excessive sweating, proptosis, goitre, pigmentation, brisk reflexes, wasting/weakness muscles	Negative	55 U/mL	>5 ng/mL	>200 ng/mL	Thyrotoxicosis	Partial thyroidectomy
		(b) 48/F	Depression	800 mg (reduced to 600 mg due to tremors)	0.9–1.28, 0.71 at 600 mg	Goitre bruit, exophthalmos, proximal myopathy, brisk reflexes, agitated, paranoid	Positive	Not reported	Not reported	Not reported	Thyrotoxicosis; eventually hypothyroid	Li reduction, carbimazole, propranolol
7	Pallisgaard and Frederiksen (1978)	50/F	BPAD	900 mg	Not reported	Enlarged non-tender goitre, fatigue, anxiety, palpitations, diarrhoea, weight loss, hyperhidrosis	Not reported	Not reported	Not reported	11.9 ug/dL (3.7–6.0)	Thyrotoxicosis (non-specific thyroiditis); euthyroid after therapy	Methimazole, Li continued, Al-triiodothyronine
8	Reus et al. (1979)	22/F	BPAD	1500 mg	0.9–1.0 mmol	Mania, heat intolerance, tremor, anxiety, palpitations	Present	Not reported	284ng100ml	16.5 ug/100 mL	Thyrotoxic self-resolved; repeat episode 1 year later; similar circumstances; transient thyrotoxicosis (silent thyroiditis) related to Li therapy	Li continued at 1500 mg
9	Fauerholdt and Vendsborg (1981)	63/M	BPAD	Nil dose, 3 year duration	Not reported	Developed goitre after 3 years and clinical and biochemical signs of thyrotoxicosis	Not reported	Not reported	Not reported	Not reported	Thyrotoxicosis	2 months antithyroid drug and partial thyroidectomy
10	Rabin and Evans (1981)	62/M	BPAD	600–900 mg	0.3–1.0	Weight loss, heat intolerance, stare with bilateral lid lag, proptosis	Not reported	5.6 uU	Normal	17.2 ug/dL (4.5–11.5)	Euthyroid initially; hyperthyroid with Li; euthyroid on cessation	Cessation of Li
11	Valenta et al. (1981)	49/F	BPAD	600 mg	0.4–0.6 mEq/L	Hyperhidrosis, fine tremor, brisk reflexes	Antimicrosomal Ab positive, Antithyroglobulin Ab negative	Blunted TRH response	270–571 ng/dL (80–190)	18.7–27.6 ug/dL	Euthyroid; hyperthyroidism post Li; euthyroid	Propylthiouracil
12	McDermott et al. (1986)	57/F	BPAD	900 mg (8 years)	0.7 mmol/L	Tachycardia, lid lag, diffuse goitre	Antithyroglobulin positive, antimicrosomal negative	3.6 uU/mL	Not reported	Elevated = 14.7–16.9 ug/dL, Free thyroxine index = 6.5–7.9	1. Primary hypothyroidism2. Hyperthyroid3. Hypothyroid (spontaneous change from Hashimoto’s to Graves’ disease)	Radioiodine; Remained on Li
13	Humphreys and Waddell (1988)	50/M	Depression	1250 mg	Not reported	Headache, chest pain, thirst, weight loss, sweating, fine tremor, palpable goitre, tachycardia brisk reflexes	Not reported	Not reported	Not reported	Recorded as elevated	1. Thyrotoxicosis2. Hypothyroid – post treatment	Radioiodine and carbimazole; Li not ceased
14	Yassa et al. (1988)	(a) 68/F	Not reported	3 years	Not reported	No clinical evidence of hypothyroidism	Not reported	0.5 uU/mL	Not reported	14.6 mcg/100 mL	Euthyroid after 2 months; transient hyperthyroidism	Li not ceased; no intervention
		(b) 61/F	Not reported	Not reported	Not reported	Exophthalmos, lid lag, warm skin, diffuse nodular goitre	Negative	Not reported	Not reported	16.5 mcg/100 mL	Thyrotoxic nodular goitre; euthyroid	Li ceased, radioactive iodine
15	Sinnott et al. (1992)	42/F	BPAD	750 mg (2 years)	0.6 mmol	Tremor, proximal muscle weakness, mildly tender diffuse goitre	Antimicrosomal Abs positive, antithyroglobulin Abs negative	Un-detectable	10.7 pmol/L (2.5–7.5)	40 pmol/L (12–28)	1. Thyrotoxicosis2. HypothyroidismLi continued; consistent with granulomatous thyroiditis	Li continued
16	Byrne and Delaney (1993)	63/M	BPAD	9 years900 mg – reduced to 600 mg after first episode hyperthyroidism. Poorly compliant – ceased	Not reported	Exophthalmos, lid lag, unable to close right eye	Not reported	Not reported	Not reported	Not reported	Hyperthyroidism – treated with carbimazole, hypothyroid – treated with thyroxine; second episode hyperthyroidism – ceased Li. Marked improvement within 72 hours – resolved	Carbimazole; Li ceased
17	Chow et al. (1993)	(a) 40/F	BPAD	1250 mg (3 years)	0.7 mmol/L	Clinically euthyroid but small diffuse goitre	Not reported	Not reported	12.9 pmol/L	Total = 173 nmol/L, FT4 = 40.9 pmol/L	1. Transient thyrotoxicosis2. transient hypothyroidism3. Euthyroidism	Li continued
		(b) 37/F	Recurrent depression	Not reported (12 year hx Li therapy)			Positive antimicoglobulin Ab, negative antisomal Ab	<0.02 ulU/L	7 pmol/L (normal)	27.5 pmol/L (elevated)	1. Hypothyroidism2. Euthyroidism on thyroxine3. Transient thyrotoxicosis after stopping thyroxine4. Hypothyroidism5. Euthyroidism	Li continued
		(c) 22/F	BPAD	750 mg	0.45	Hand tremor, heat intolerance, weight loss	Negative	<0.02 ulU/L	16.4 pmol/L	>67 pmol/L	1. Transient thyrotoxicosis2. Transient hypothyroidism3. Euthyroidism	Li continued
18	Becerra-Fernandez (1995)	48/F	BPAD	2 years, 600 mg	1.12 mEq/L (0.2–1.2)	Diffuse goitre	Negative	0.06 uIU/mL (0.32–5.56)	3.34 ng/dL (0.8–2.67)	283 ng/dL (80–180)	Hashimoto’s disease; Li ceased; propranolol and carbimazole; euthyroid, antibodies remained undetectable; Li-induced or underlying disease, but antibody negative suggests role of Li	Li ceased; propranolol; carbimazole
19	Mizukami et al. (1995)	26/F	BPAD	800 mg, 2 year duration	Not reported	Sweating, palpitations, anorexia, diffusely enlarged goitre, no pre-existing thyroid dysfunction	Negative (TG and M)	<0.2 uU/L	FT3 = 13.6 pmol/L (3.4–8.2)	FT4 = 87.5 pmol/L	Li-associated autoimmune thyroiditis	Li ceased; propranolol
20	Shimzu et al. (1997)	29/F	BPAD	600 mg, 27 month duration	Not reported	Palpitations, irritation, weight loss, no pre-existing thyroid dysfunction	Positive TSH binding inhibitor Ig	<0.05 uIU/mL (0.03–5.00)	3.6 nmol/L (1.2–3.1)	233 nmol/L (58–155)	Euthyroid initially; developed Graves’ disease; euthyroid with treatment	Methimazole; eventually thyroidectomy
21	Dwarakanathan (1998)	(a) 31/F	BPAD – depression	900 mg	Not reported	Tachycardia, anxiety, diffusely enlarged thyroid	Not reported	>0.2 uU/mL (0.71–7.5)	Not reported	15.4 uU/dL (4.4–10.7)	Li-induced hyperthyroidism	Propylthiouracil PTU therapy 100 mg TDS, radioactive iodine
		(b) 48/F	Depression	7 months	Not reported	Tremor, weight loss, heat intolerance, anxiety, palpitations, diffusely enlarged thyroid	Not reported	>0.03 uU/mL (0.38–0.615)	Not reported	19 uU/dL	Li-induced hyperthyroidism	Radioactive iodine
		(c) 50/F	Depression	17 years	Not reported	Itching, gastrointestinal symptoms, tremor, anxiety, diffusely enlarged thyroid	Not reported	>0.03 uU/mL (0.38–0.615)	Not reported	24 uU/dL (4.4–10.7	Li-induced hyperthyroidism	Radioactive iodine
22	Yamagishi and Yokoyama-ohta (1999)	50/F	BPAD	800 mg daily, 5 month duration	1.0 mmol	Symptomatic 3 months	Negative	<0.03 uIU/ml (0.4–5.0)	27.10 pmol/L (3.39–6.31)	81.72 pmol/L (10.3–21.88)	Pre-Li TFTs normal; Graves’ disease	Not reported
23	Onuigbo et al. (2000)	47/M	BPAD	1200 mg, 10 year duration	1.9 mmol/L	Tremor, ataxia, tachycardia, increased reflexes	Negative	<0.06 uU/mL (0.38–4.7)	Increased T3 uptake	Free T4 = 5.7nh/dL (2.3–4.2)	Euthyroid pre-Li; painless thyroiditis (Li-associated thyrotoxicosis)	Atenolol; Li ceased
24	Dang and Hershman (2002)	52/M	BPAD	900 mg for 15 years, ceased 3 months before admission	>0.1	Delusional mania, anxious, weight loss, tachycardia	Negative	0.06 uU/mL (0.6–4.6)	Not reported	6.17 ng/dL (0.8–2)	Silent thyroiditis – atypical granulomatous thyroiditis	Atenolol, prednisone; ceased Li
25	Numata et al. (2002)	48/F	Depression	600 mg for 9 days	0.55–0.63 mEq/L	Perspiration	Not detectable	0.01 ng/dL (0.65–5.55)	6.5 ng/dL (2.3–3.7)	2.4 ng/dL (0.93–1.75)	Thyrotoxicosis caused by silent thyroiditis	Lithium ceased and thyroid function normalised
26	Law et al. (2004)	39/M	BPAD	No dose recorded, 18 year duration	Therapeutic	Weight loss, tachycardia, tremor	Not reported	<0.1 (0.4–5.0 uIU/L)	7.2 (0.7–2.1 nmol/L)	53 pmol/L (10–23)	Li associated thyrotoxicosis; normalised with treatment	Propranolol
27	Dalan et al. (2007)	65/F	BPAD	Not reported	0.6	Delirious, restless dehydrated, multinodular goitre	Negative	0.05 uIU/L (0.29–0.377)	415.6 pg/dL (279.2–538.96)	1.24 ng/dL	Subclinical hyperthyroidism	Ceased Li, carbimazole
28	Arlt et al. (2008)	56/F	BPAD	4 months, 1125 mg	1.0	After iodine contrast medium, weakness, numbness, gait disturbance, weight loss, depressed mood, progressed to tetraparesis, bulbar symptoms, disorientation, paranoid/ nihilistic delusions and cachexia	Negative	0.47 uU/L(0.35–4.5) – dropped less than 0.01	Normal on admission, 222 ng/dL	Normal on admission, 8.1 then increased to 14.7 ug/dL (4.5–11.5)	Thyrotoxicosis	Thiamazole and prednisolone; ceased Li and symptoms largely resolved in 3 weeks
29	Bernstein and Friedman (2011)	61/M	BPAD	600 mg for 7 years	0.5 mmol/L	Not symptomatic, no pre-existing thyroid illness	Negative	0.03 uU/L (0.34–5.60)	4.8 pg/mL (2.4–4.2)	1.73 ng/dL (0.58–1.64)	Thyroiditis	Lithium tapered and ceased – 4 weeks later TFTs normalised
30	Deardorff et al. (2016)	38/F	Schizoaffective	900 mg	0.8 mmol/L	Akathisia, weight loss	Not reported	0.004 mIU/L	Not reported	1.48 mIU/L	Li-induced hyperthyroidism	Beta blockers; thyroid status normalised after 50 days from Li cessation
31	Bandyopadhyay and Nielsen (2012)	46/F	BPAD	Not reported	3.6mEq/L (0.6–1.2), Li toxic	Acute onset agitation and delirium, hyperhidrosis, dehydrated, tachycardia, bilateral pedal oedema, coarse tremor, brisk reflexes, no goitre	Not reported	0.016 uU/mL (0.4–5.5)	7.5 pg/mL (1.8–4.6)	3.0 ng/dL (0.7–1.8)	Severe thyrotoxicosis and impending thyroid storm	Haemodialysis; propylthiouracil; propranolol
32	Sato et al. (2013)	64/F	BPAD	Maintenance dose for 9 years	Therapeutic level	Mild diarrhoea, malaise and low-grade fever, drowsiness, proximal myopathy	Antithyroid Ab negative, peroxidase Ab positive	<0.01	8.12 pg/mL (1.71–3.71)	4.45 ng/dL (0.7–148)	Thyrotoxicosis due to silent thyroiditis	Li ceased
33	Siyam et al. (2013)	27/M	Schizoaffective disorder (3 years hx)	1350 mg		Denied diaphoresis, heat intolerance, eye protrusion, tremor, weight gain, tachycardia, AH/VH	Negative	Low = 0.006 mcU/mL (0.270–4.2)	4.6 pg/mL (2.0–4.4)	1.57 ng/dL (90.93–1.70)	Acute thyroiditis; hyperthyroidism resolved during admission, 2 month post – complete normalisation TFT	Li discontinued on admission
34	Tan et al. (2013)	51/F	BPAD (18 years hx)	1350 mg	0.4–0.6 mmol/L	Symptomatic, no pre-existing thyroid illness	Not reported	0.02 mIU/L (0.5–4.0)	Not reported	19 pmol/L (10–25)	Hyperthyroidism due to silent thyroiditis	Carbimazole; Li continued; symptoms resolved after 3 months then hypothyroidism
35	Chalasani and Benson (2014)	19/F	BPAD – subclinical hypothyroidism	1250 mg	0.75 mmol/L	2/52 hx headaches, lethargy, weight loss, tremor, achycardia, diffuse non-tender goitre, lid lag	Negative	0.01 mU/L (0.4–4.0)	Not reported	63.6 pmol/L (12–22)	Hyperthyroidism improved when propranolol commenced, lithium continued, hypothyroid result	Li continued; propranolol
36	El-Bakush et al. (2014)	25/F	BPAD – schizophrenia, pre-existing hypothyroidism	Not reported, recently increased	4.6 mmol/L	Lithium toxicity	Antithyroid peroxidase Ab positive	<0.002 mU/mL (low)	Not reported	1.96 ng/dL (elevated)	Lithium toxicity, silent thyroiditis	Propylthiouracil; haemodialysis
37	Kar et al. (2014)	61/F	BPAD (23 years hx)	18 year duration on lithium	Therapeutic range at onset thyrotoxicosis however fluctuated up to 1.2 mmol/L	Thyrotoxic, no previous history of thyroid illness	Not reported	0.01 mU/L (0.27–4.20 mU/L)	Not reported	Elevated series of 3 months = 57, 72 and 40 pmol/L (12.0–22.0)	Thyrotoxicosis; after 5 months treatment, developed hypothyroidism	Li continued and titrated according to serum Li levels; carbimazole
38	Altieri et al. (2015)	Adult female	BPAD – recent sleeve gastrectomy	Not reported	2.7 mEq/L (0.5–1.2), Li Toxic	Tremors, reduced oral intake, emesis, agitated, tachycardia hypertensive, febrile	Not reported	Low	Elevated	Elevated	Thyroid storm; TFTs normalised after treatment	Methimazole; propranolol
39	De Sousa Gurgel et al. (2015)	19/M	Schizoaffective disorder	1200 mg	0.61 mmol/L	Anxiety headaches, restlessness that progressed to severe psycho-motor agitation, tachycardia, tremor	Not reported	0.61 uUI/mL (0.34–5.6)	Not reported	FT4 = 3.13 ng/dL (0.54–1.24) elevated	Thyrotoxicosis	When Li was ceased, symptoms rapidly resolved

TSH: thyroid stimulating hormone; BPAD: bipolar affective disorder; TFT: thyroid function test.

Lithium was used for a range of diagnoses including bipolar disorder (n = 30 patients), depression (n = 10 patients), schizoaffective disorder (n = 2 patients) and cyclothymic personality disorder (n = 1 patient). Diagnoses were missing for two patients. The mean age at presentation was 46 (standard deviation = 14, range = 19–68) years, with the majority female (66%). The median lithium dose was 900 (range = 600–1500) mg/day, and the mean and median duration of lithium therapy before developing thyrotoxicosis was 3 years (range = 9 days to 18 years). Almost all patients (44 of 45) were biochemically and clinically hyperthyroid with only one case being chemically hyperthyroid but clinically asymptomatic. Autoantibodies were positive in 9 cases, negative in 15 and not reported in 21 cases.

The three case series described 46 patients with some form of hyperthyroidism while on lithium treatment (Table 3) (Barclay et al., 1994; Brownlie and Turner, 2011; Sirota et al., 1992). In terms of socio-demographic features, 36 were females and 10 males, with a mean age of 45 years. The mean duration of treatment was 5.3 (range = 0.5–25) years. Thyroid illness was treated in 24 patients (52%) and not treated in 9 (19.6%), with no information on treatment in 13 patients (28.3%).

Table 3.

Case series studies.

Author(s), year	Study design and recruitment	Psychiatric diagnosis	Outcomes measures	No. of participants	Age/gender	Duration of lithium therapy	Findings
Brownlie and Turner (2011)	Study of hyperthyroid patients presenting to clinic who were maintained on Li therapy between 1999 and 2006	Thyrotoxicosis; psychiatric diagnosis not reported	TFTs, 99mTC pertechtenate thyroid scintiscan	23	20 F, 3 M; median age = 41 (19–75) years	6 (0.6–25) years	10 diffuse hyperplasia – 2 of which Graves’ disease, 2 toxic multi-nodular goitre, 9 painless thyroiditis, 1 sub-acute thyroiditis (on Li for <1 year), 1 diffuse goitre
Barclay et al. (1994)	Retrospective study identifying 14 pts with Li-associated thyrotoxicosis attending thyroid clinic over 18-year period (1973–1991)		Tc pertechtenate thyroid scans, serum T3, T4, free T4 index, thyroid microsomal and thyroglobulin Ab titres	14	9 F, 5 M; mean age = 49 (17–72) years	4.3 (1–11) yearsMean Li dose = 750 (400–1250) mg	Scintiscans of 13 pts – identified 8 toxic diffuse goitre, 2 toxic multi-nodular goitre, 1 toxic uninodular goitre, 2 painless thyroiditis
Sirota et al. (1992)	Case series report of hyperthyroidism in pts on Lithium	Psychiatric illness	TFTs, physical stigmata	9	7 F, 2 M; average age = 50 years	Average duration = 59 months (6 months–15 years)	All 9 had goitres, 8 eye findings, 7 clinical Graves’ disease, 1 toxic multi-nodular goitre, 1 solitary toxic nodule

TFT: thyroid function test.

There were nine cross-sectional or cohort studies (Table 4) that described 2674 patients who were treated with lithium (Bocchetta et al., 2007a, 2007b; Calabrese et al., 1985; Caykoylu et al., 2002; Hayes et al., 2016; Kirov, 1998; Kirov et al., 2005; Kuman Tuncel et al., 2017; Shine et al., 2015). Of these patients, 58% were females. The mean age was 46.5 years, and the mean duration of lithium therapy was 2.8 years. Of these, 2.6% (n = 72) were hyperthyroid. Only one study reported the mean lithium dose (Kuman Tuncel et al., 2017). One further large study of laboratory results (n = 235, 951) was not included in the descriptive statistics in the absence of usable raw data and measures of either lithium duration or dosage (Shine et al., 2015). In this study, lithium levels were not associated with hyperthyroidism after adjustment for age, sex and comorbidity (Shine et al., 2015). However, there was an increased risk in patients with lithium levels that were higher than the median for the sample (0·6 mmol/L).

Table 4.

Cross-sectional and cohort studies.

Author(s), year	Study design and recruitment	Psychiatric diagnosis	Outcome measures	No. of participants	Gender/age	Duration of lithium therapy	Findings
Bocchetta et al. (2007b)	Prospective study of clinically relevant thyroid dysfunction in a cohort of patients on long-term lithium treatment	Not stated	At least one TSH level, thyroid circulating antibodies, hypothyroidism/ hyperthyroidism requiring treatment	118 out of 150 pts on lithium	66% F at baseline; median age = 40.0 years: F, 39.0 years, M at baseline	Up to 15 years; mean = 8.3 years	1 case of hyperthyroidism = an annual incidence rate of 0.1%
Bocchetta et al. (2007a)	Prospective comparison of thyroid and renal function of patients who had started on lithium before and after 65 years of age	BPAD1, BPAD2 or major depression	Hyperthyroidism requiring antithyroid medication	Of 110 pts, 58 still on lithium, 20 died, 32 lost to follow-up	73% F at baseline.	6 years; age range = 65–84 years	2 new cases of hyperthyroidism
Kuman Tuncel et al. (2017)	Cross-sectional study design with patients recruited from an affective disorders outpatient unit and controls recruited from hospital staff	BPAD	Thyroid hormone antibodies, thyroid ultra sonography	84 patients treated with lithium matched and 65 gender and age similar controls	53.6% F; mean age = 43.1 years	Mean duration of lithium therapy = 139.4 months; mean dose = 1088.8 mg/day	1.2% of lithium pts (1 case) of subclinical hyperthyroidism.No statistical difference between cases and controls.
Hayes et al. (2016)	Propensity score adjusted longitudinal cohort study using nationally representative UK electronic health records from January 1995 to December 2013 of pts on olanzapine, quetiapine, valproate and lithium	BPAD1	Identified adverse outcomes including hyperthyroidism, TSH < 0.1 mU/L	2148 pts on lithium	59.92% F; median age = 46.28 years	Pts were included if they had received at least a 1-month prescription of lithium	41 events of hyperthyroidism.Rate per 100 patient year at risk (PYAR) = 0.78.Rate increased compared to valproate (p = 0.003) and olanzapine (p = 0.007) but not quetiapine (p = 0.096).
Kirov et al. (2005)	Cohort study.Patients were recruited from a lithium clinic or as part of a genetic association study with a 14–87 month follow-up period.	BPAD1, BPAD2 or unipolar depression	TFTs, Ab assays	57	33 F, 24 M; mean age = 49.3 (24–87) years	Mean = 53.1 (14–87) months	1 thyrotoxicosis (Graves’ disease)
Caykoylu et al. (2002)	Cross-sectional study of clinic patients with diagnosis of BPAD and on lithium.Exclusion of previous thyroid dysfunction or augmenting psychotropic agents.	BPAD1	TFTs	42	17 F, 25 M; mean age: F, 36.7 ± 15.3 years; M, 29.1 ± 11.7 years	Duration range = 4–156 months; dose between 600 and 1200 mg/day	3 M subclinical hypothyroidism; 3 F subclinical hyperthyroidism; 1 M hyperthyroidism; toxic goitre 16 pts (10 F)
Kirov (1998)	Retrospective study design of patient records within the catchment area of the Maudsley Hospital, London. Includes a short prospective arm of 3–20 months	BPAD1/2, MDD, schizoaffective disorder (SCAD)	TFTs	Total 209: 108 BPAD1, 19 BPAD2, 63 recurrent unipolar MDD, 19 SCAD	Caucasian; 87 M, 122 F; mean age = 52.8 (22–94) years	6.4 years	2 (1 M and 1 F) developed thyrotoxicosis
Calabrese et al. (1985)	Prospective cohort study of lithium naïve outpatients with 4–20 month period of follow-up	BPAD (DSM-III)	TFTs, Ab assays every 4 months from 4 to 20 months	16	10 F, 6 M; mean age = 44 (17–61) years	14.5 ± 4 months	1 patient with pre-existing positive thyroid microsomal antibodies developed hyperthyroidism
Shine et al. (2015)	Retrospective cohort study of routinely collected data from laboratory information systems between 1 October 1985 and 31 March 2014	Unable to be determined due to source of information collection	At least two measurements of serum creatinine, thyrotropin, calcium, HbA1c or lithium level	1916 pts with serum lithium levels matched with 234,035 non lithium exposed controls	1052 F, mean age = 49 (35–66) years; 864 M, mean age = 49 (37–61) years	Unable to be determined due to source of information collection	Lithium not associated with hyperthyroidism based on suppression of thyrotropin activity (p = 0.1010)

TSH: thyroid stimulating hormone; BPAD: bipolar affective disorder; MDD: major depressive disorder; DSM-III: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition; TFT: thyroid function test.

There was only one case control study of 400 patients referred to Massachusetts General Hospital, 30 of whom had Graves’ disease and 100 silent thyroiditis (Miller and Daniels, 2001). The odds of lithium exposure was increased 4–7 fold in patients with silent thyroiditis compared to those with Graves’ disease (95% confidence interval [CI] = [1.3, 17.1]).

The majority of the studies included involved very small numbers of participants, with only four studies having more than 200 participants (Hayes et al., 2016; Kirov, 1998; Miller and Daniels, 2001; Shine et al., 2015). However, two further studies had large numbers of patients treated with lithium (n = 1916 with 234,035 matched controls and n = 2148 patients, respectively) (Hayes et al., 2016; Shine et al., 2015). Length of follow-up was highly variable, both within and between studies, and in a single study was inadequately defined (Kirov, 1998). Inclusion and exclusion criteria also differed between studies and were not stated in two studies (Calabrese et al., 1985; Kirov et al., 2005). However, they were more extensive and better elaborated in more recent studies (Caykoylu et al., 2002; Hayes et al., 2016; Kuman Tuncel et al., 2017). Three studies took into account possible confounders in their statistical analysis (Hayes et al., 2016; Kuman Tuncel et al., 2017; Shine et al., 2015).

In the case of cross-sectional, case control or cohort studies where we were able to use the modified Newcastle–Ottawa scale, 8 of the 10 studies were judged as being at low risk of bias (Bocchetta et al., 2007a, 2007b; Caykoylu et al., 2002; Hayes et al., 2016; Kirov, 1998; Kuman Tuncel et al., 2017; Miller and Daniels, 2001; Shine et al., 2015) (Table 5).

Table 5.

Modified Newcastle–Ottawa scale.

Study	Representativeness	Size	Comparability	Outcome	Statistics	Total
Bocchetta et al. (2007b)	0	0	1	1	1	3
Bocchetta et al. (2007a)	0	0	1	1	1	3
Kuman Tuncel et al. (2017)	0	0	1	1	1	3
Hayes et al. (2016)	1	1	1	1	1	5
Shine et al. (2015)	1	1	1	1	0	4
Kirov et al. (2005)	0	1	0	1	0	2
Caykoylu et al. (2002)	0	0	1	1	1	3
Miller and Daniels (2001)	0	1	1	1	1	4
Kirov (1998)	0	1	0	1	1	3
Calabrese et al. (1985)	0	0	0	1	0	1

Discussion

Lithium therapy is commonly associated with hypothyroidism. By contrast, hyperthyroidism is a very low prevalence adverse event associated with lithium therapy. To our knowledge, this is the first systematic review of all published reports of hyperthyroidism associated with lithium therapy. We found 52 studies, 39 of which were individual case reports and 3 case series. There were 10 cross-sectional, case control or cohort studies.

The mechanisms by which lithium induces a hyperthyroid state are unclear, with several different theories proposed by researchers. One explanation is an underlying autoimmune mechanism, especially in patients with diffuse hyperplasia of the thyroid (Dang and Hershman, 2002). For instance, lithium leads to increased immunoglobulin production in in-vitro studies (Deardorff et al., 2016). There is also an increase in B-cell activity and a decrease in ratios of cytotoxic T-cells among lithium-treated patients, as well as increases in antithyroid antibodies (either antimicrosomal or antithyroglobulin) among lithium-treated patients with affective disorders (20%) compared to those not on lithium (7.5%) (Wilson et al., 1991). A further study found that lithium increased the titre of antithyroid antibodies in patients who already possessed them, but not in control subjects exposed to lithium, suggesting that at the very least that lithium can exacerbate pre-existing thyroid autoimmunity (Rapaport, 1994). Finally, autoimmunity is suggested by the occurrence of exophthalmos among lithium-treated patients, which resolves on cessation (El-Bakush et al., 2014).

However, an autoimmune process cannot be the sole explanation given that antibodies were positive in less than 50% of patients in the case reports we identified where it was measured (Table 2). Another possible mechanism is iodide retention and expansion of intra-thyroidal iodide stores secondary to inhibition of thyroid hormone synthesis and release. This increase in intra-thyroidal iodide may precipitate thyrotoxicosis in patients who are genetically predisposed to Graves’ disease, or nodular goitres, and thus particularly sensitive to the effects of iodide (Barclay et al., 1994; McDermott et al., 1986). Finally, lithium may exert a direct toxic effect on the thyroid gland (Altieri et al., 2015; Humphreys and Waddell, 1988), although evidence is weaker given the low numbers of patients involved and the absence of any histological evidence of direct toxic changes in their thyroid gland cells (Miller and Daniels, 2001).

The time course from first exposure to lithium to the development of hyperthyroidism is highly variable, ranging from weeks to years. Thus, lithium can have an immediate or a delayed effect, and the reason for this is unclear. It may be that when hyperthyroidism occurs shortly after the introduction of lithium, there is a direct toxic reaction or destructive thyroiditis, akin to amiodarone-induced toxicity type II, while a delay of months or years suggests an autoimmune process (Siyam et al., 2013).

Clinical significance

There are a number of important clinical considerations. For instance, sudden changes in thyroid hormone levels may precipitate a psychiatric relapse (Arlt et al., 2008), while thyrotoxicosis in a patient with an affective disorder may mimic the symptoms of their psychiatric illness and therefore be easily missed (De Sousa Gurgel et al., 2015). This underscores the importance of checking thyroid function in acutely ill psychiatric patients, particularly if they are on lithium therapy and not responding to treatment.

Furthermore, hyperthyroidism may also alter the renal clearance of lithium, leading to toxicity. Four cases of lithium toxicity in patients with thyrotoxicosis have been reported (Bandyopadhyay and Nielsen, 2012; Oakley et al., 2000; Sato et al., 2013). All required haemodialysis to treat their lithium toxicity, and one experienced serious complications of cardiac arrest, quadriplegia and coma, although she ultimately made a full recovery (Sato et al., 2013). It is possible that in hyperthyroid patients, altered tubular function due to induction of sodium–hydrogen counter-transport increases proximal tubule absorption of lithium, while simultaneously reducing the fractional excretion of lithium (Oakley et al., 2000). It is therefore important for clinicians to be aware that not only patients receiving lithium therapy may be at an increased risk of thyrotoxicosis but that this state can also induce lithium toxicity. In patients where there is a suspicion of lithium toxicity, thyroid function tests and lithium level should therefore be performed as a routine part of the initial evaluation (Sato et al., 2013).

Silent thyroiditis is characterised by thyrotoxicosis with a very low 24-hour radioiodine uptake, and spontaneously resolves, but may be followed by a hypothyroid phase (Woolf, 1980). In most cases, it is detected on routine testing due to its transient nature and less obvious clinical features when compared to Graves’ disease (Brownlie and Turner, 2011). For this reason, silent thyroiditis is more likely to elude diagnosis altogether. It is also possible that in some patients diagnosed with lithium-induced hypothyroidism, what is actually being identified is the hypothyroid phase of silent thyroiditis, where the thyrotoxic phase has been missed (Miller and Daniels, 2001).

By contrast, Graves’ disease is more likely to present symptomatically and the thyrotoxicosis is typically characterised by a diffuse goitre, elevated radioiodine uptake and homogeneous findings on scanning (McDermott et al., 1986). As it is an autoimmune disorder, it is unclear what role, if, any lithium plays in the pathogenesis of this condition in lithium-treated patients. Graves’ disease is a relatively common condition, and therefore, a proportion of lithium-treated patients will develop the disease coincidentally to their lithium treatment (Thompson and Baylis, 1986). However, as discussed above, there are researchers who suggest a link between lithium therapy and autoimmunity, either by precipitating illness in an already predisposed individual or perhaps by a direct effect (Fauerholdt and Vendsborg, 1981; Miller and Daniels, 2001). Interestingly, the case reports included four cases of ‘Hashitoxicosis’, three described by Chow et al. (1993) (one of these was not included in our table as it followed lithium discontinuation) and one described by McDermott and colleagues (Chow et al., 1993; McDermott et al., 1986). ‘Hashitoxicosis’ is used to describe the phenomenon whereby Hashimoto’s thyroiditis (with or without hypothyroidism) spontaneously evolves into Graves’ disease associated with hyperthyroidism. It is considered rare, leading Chow et al. (1993) to argue that the association with lithium therapy may be at least partly causal (Chow et al., 1993).

Our findings on hyperthyroidism highlight the importance of monitoring lithium levels closely both during initiation and on-going management. For this and other reasons, lithium levels should be as low as possible while consistent with clinical gain given findings that higher levels are associated with an increased risk of hyperthyroidism. Clinical practice guidelines have generally converged around recommended lithium levels of between 0.6 and 0.8 mmol/L as maintenance therapy (Malhi et al., 2016). Given that even at this level, there is the possibility of increased risk (Shine et al., 2015), thyroid function tests and, if possible, thyroid antibodies should be measured at initiation and every subsequent 6 months (McKnight et al., 2017). The ‘Lithiumeter’ was developed as a visual guide for determining lithium levels in the management of bipolar disorder and has recently been updated (Malhi et al., 2016).

The treatment of hyperthyroidism, if it does occur, depends on the aetiology. As patients with silent thyroiditis are usually asymptomatic, it is appropriate not to institute any therapy, but to perform serial monitoring of thyroid function, as greater than 50% will subsequently develop hypothyroidism (Brownlie and Turner, 2011). As the course of silent thyroiditis (spontaneous remission, followed by either euthyroid or hypothyroid state) seems to be independent of the decision to continue or withdraw lithium therapy, it would seem reasonable to continue lithium therapy in those patients for whom it has been helpful maintaining remission of their psychiatric symptoms (Chalasani and Benson, 2014). The natural history of ‘Hashitoxicosis’ is also transient and self-limiting, so antithyroid therapy may again not be necessary (Chow et al., 1993).

When indicated, antithyroid drugs such as carbimazole, with or without steroids, are effective in cases of lithium-induced hyperthyroidism (Brownlie and Turner, 2011). For patients with lithium-induced Graves’ disease, radioiodine or thyroidectomy may also be necessary, particularly if patients are non-compliant with antithyroid drugs. In cases of toxic nodular goitre, even in the absence of compressive symptoms, thyroidectomy is indicated (Brownlie and Turner, 2011). Common interventions in our case reports included antithyroid agents, and the reduction or cessation of lithium. Three required surgery. Based on the case studies reported in this review, there is no universal recommendation for either continuing or ceasing lithium therapy after hyperthyroidism is identified, although in recent years discontinuing lithium has been the most commonly reported.

Strengths and limitations

By systematically reviewing the literature, we report the largest collection to date of case reports, case series, cross-sectional, cohort and case control studies on the association between lithium therapy and hyperthyroidism. However, there are several limitations. Many of the papers were of variable quality as most were case studies or series of small numbers of participants rather than cross-sectional, case control or cohort studies. While most of the latter designs had a low risk of bias, only one measured serum lithium. Psychiatric diagnoses, lithium dose and level, and length of treatment were inconsistently reported. There was heterogeneity in the tests used to determine thyroid illness reflecting changes in practice over time. In addition, thyroid assays have become more sensitive and reference ranges for the diagnosis of illness have altered, leading to an increase in rates of detection (Kuman Tuncel et al., 2017). A clear diagnosis of thyroid illness was also not always available.

Furthermore, not all studies reported thyroid function prior to commencing lithium, so pre-existing thyroid illness could not be excluded thereby raising questions as to whether a temporal relationship exists in these cases. The limited number of prospective studies meant that a clear temporal relationship could not be established and there was no evidence of a dose response. A plausible alternative explanation for any association might be that patients on lithium are more likely to have their thyroid function monitored as per clinical guidelines and thus incident cases are more likely to be detected (Malhi et al., 2015). There was also limited evidence of a dose response although it is worth noting that the largest of the included papers (n = 235, 951) reported an increased risk of hyperthyroidism at higher lithium levels that was not present at concentrations below the study median (Shine et al., 2015).

In addition, concomitant psychotropic medication was seldom reported and it is possible that other medication could be responsible for changes in thyroid function. Indeed, one study found that rates of hyperthyroidism increased in people prescribed lithium compared to valproate and olanzapine, but not quetiapine (Hayes et al., 2016). Together, these make it difficult to make any conclusions regarding the nature of the association between lithium treatment and hyperthyroidism. Large prospective studies are required to clarify this and to inform clinical management of patients treated with lithium where hyperthyroidism occurs.

Finally, our search was limited to studies in English and we were unable to undertake a meta-analysis because of the heterogeneity of study populations and designs.

Conclusion

Hyperthyroidism is uncommon in patients on lithium compared to hypothyroidism. Nonetheless, it has several clinical implications. In the acute setting, it can complicate or mimic mania and lead to lithium toxicity. Distinguishing a relapse of bipolar illness from thyroid dysfunction is critical, as it determines the most appropriate treatment and hence patient outcome. Given the high prevalence of any thyroid illness in patients receiving lithium and the significant clinical complications, it is essential that all patients requiring mood stabilisation be advised of the potential adverse effects of lithium so they can make informed choices about their health care. Regular thyroid monitoring should occur in all patients prescribed lithium.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

J.G.S. is supported by a National Health and Medical Research Council Practitioner Fellowship Grant (grant no. 1105807).

ORCID iD’S

James G Scott

Steve Kisely

References

Altieri

Pryor

Vosswinkel

et al . (2015) Lithium induced thyroid storm after sleeve gastrectomy. Critical Care Medicine 43: 316.

Arlt

Burkhardt

Wiedemann

(2008) Thyrotoxicosis after iodine contrast medium administration: Rapid mood swing to mania and subsequent psychotic depression in a patient with bipolar disorder during lithium therapy. Pharmacopsychiatry 41: 163–165.

Bafaqeeh

Myers

(1976) Letter: Lithium and thyrotoxicosis. The Lancet 1: 1409.

Bandyopadhyay

Nielsen

(2012) Lithium-induced hyperthyroidism, thyrotoxicosis and mania: A case report. QJM: Monthly Journal of the Association of Physicians 105: 83–85.

Barclay

Brownlie

Turner

et al . (1994) Lithium associated thyrotoxicosis: A report of 14 cases, with statistical analysis of incidence. Clinical Endocrinology 40: 759–764.

Becerra-Fernandez

(1995) Autoimmune thyrotoxicosis during lithium therapy in a patient with manic-depressive illness. American Journal of Medicine 99: 575.

Bernstein

Friedman

(2011) Lithium-associated hyperthyroidism treated with lithium withdrawal: A case report. American Journal of Psychiatry 168: 438–439.

Bocchetta

Loviselli

(2006) Lithium treatment and thyroid abnormalities. Clinical Practice & Epidemiology in Mental Health 2: 23.

Bocchetta

Cabras

Pinna

et al . (2007a) An observational study of 110 elderly lithium-treated patients followed up for 6 years with particular reference to renal function. International Journal of Bipolar Disorders 5: 19.

10.

Bocchetta

Cocco

Velluzzi

et al . (2007b) Fifteen-year follow-up of thyroid function in lithium patients. Journal of Endocrinological Investigation 30: 363–366.

11.

Bou Khalil

Richa

(2011) Thyroid adverse effects of psychotropic drugs: A review. Clinical Neuropharmacology 34: 248–255.

12.

Brownlie

Turner

(2011) Lithium associated thyrotoxicosis. Clinical Endocrinology 75: 402–403.

13.

Byrne

Delaney

(1993) Regression of thyrotoxic ophthalmopathy following lithium withdrawal. Canadian Journal of Psychiatry 38: 635–637.

14.

Calabrese

Gulledge

Hahn

et al . (1985) Autoimmune thyroiditis in manic-depressive patients treated with lithium. American Journal of Psychiatry 142: 1318–1321.

15.

Carmaciu

Anderson

Lawton

(2003) Thyrotoxicosis after complete or partial lithium withdrawal in two patients with bipolar affective disorder. Bipolar Disorders 5: 381–384.

16.

Caykoylu

Capoglu

Unuvar

et al . (2002) Thyroid abnormalities in lithium-treated patients with bipolar affective disorder. Journal of International Medical Research 30: 80–84.

17.

Chalasani

Benson

(2014) Lithium-induced thyrotoxicosis in a patient with treatment-resistant bipolar type I affective disorder. Medical Journal of Australia 201: 541–542.

18.

Chow

Lee

Shek

et al . (1993) Lithium-associated transient thyrotoxicosis in 4 Chinese women with autoimmune thyroiditis. Australian and New Zealand Journal of Psychiatry 27: 246–253.

19.

Cubitt

(1976) Letter: Lithium and thyrotoxicosis. The Lancet 1: 1247–1248.

20.

Dalan

Leow

Jong

(2007) Multiple endocrinopathies associated with lithium therapy. Endocrine Practice 13: 758–763.

21.

Dang

Hershman

(2002) Lithium-associated thyroiditis. Endocrine Practice 8: 232–236.

22.

De Sousa Gurgel

Dutra

Higa

et al . (2015) Hyperthyroid rage: When bipolar disorder hides the real disorder. Clinical Neuropharmacology 38: 38–39.

23.

Deardorff

Gwozdziewycz

Adusumilli

(2016) A case of mistaken identity: Akathisia or lithium-induced hyperthyroidism? Primary Care Companion for CNS Disorders 18. Available at: https://www.psychiatrist.com/PCC/article/Pages/2016/v18n05/15l01921.aspx

24.

Dekkers

Egger

Altman

et al . (2012) Distinguishing case series from cohort studies. Annals of Internal Medicine 156: 37–40.

25.

Dwarakanathan

(1998) Hyperthyroidism during lithium therapy for depression. Endocrine Practice 4: 201–203.

26.

El-Bakush

Fuller

Tello

et al . (2014) Diverse complications in a patient with lithium toxicity. Journal of Investigative Medicine 62: 430.

27.

Fauerholdt

Vendsborg

(1981) Thyroid gland morphology after lithium treatment. Acta Pathologica et Microbiologica Scandinavica. Section A, Pathology 89: 339–341.

28.

Franklin

(1974) Thyrotoxicosis developing during lithium treatment: Case report. New Zealand Medical Journal 79: 782.

29.

Hayes

Marston

Walters

et al . (2016) Adverse renal, endocrine, hepatic, and metabolic events during maintenance mood stabilizer treatment for bipolar disorder: A population-based cohort study. PLoS Medicine 13: e1002058.

30.

Humphreys

Waddell

(1988) Lithium, psoriasis, abnormal glucose tolerance, and thyroid dysfunction. British Journal of Psychiatry 152: 437–438.

31.

Iga

Taniguchi

Ohmori

(2005) Mood swing from severe depression to mania following acute alteration of thyroid status. General Hospital Psychiatry 27: 451–453.

32.

Kar

Hullumane

Williams

(2014) Thyrotoxicosis followed by hypothyroidism in a patient on lithium. Mental Illness 6: 5415.

33.

Kibirige

Luzinda

Ssekitoleko

(2013) Spectrum of lithium induced thyroid abnormalities: A current perspective. Thyroid Research 6: 3.

34.

Kirov

(1998) Thyroid disorders in lithium-treated patients. Journal of Affective Disorders 50: 33–40.

35.

Kirov

Tredget

John

et al . (2005) A cross-sectional and a prospective study of thyroid disorders in lithium-treated patients. Journal of Affective Disorders 87: 313–317.

36.

Kuman Tuncel

Akdeniz

Ozbek

et al . (2017) Thyroid function and ultrasonography abnormalities in lithium-treated bipolar patients: A cross-sectional study with healthy controls. Noro Psikiyatri Arsivi 54: 108–115.

37.

Law

Rifkind

Rosen

et al . (2004) Hyperthyroidism six years post-BMT for acute myeloblastic leukemia in a patient on long-term lithium therapy for bipolar affective disorder. Leukemia & Lymphoma 45: 807–809.

38.

Lazarus

(2009) Lithium and thyroid. Best Practice & Research: Clinical Endocrinology & Metabolism 23: 723–733.

39.

Lazarus

McGregor

Ludgate

et al . (1986) Effect of lithium carbonate therapy on thyroid immune status in manic depressive patients: A prospective study. Journal of Affective Disorders 11: 155–160.

40.

McDermott

Burman

Hofeldt

et al . (1986) Lithium-associated thyrotoxicosis. American Journal of Medicine 80: 1245–1248.

41.

MacGregor

(1977) Hyperthyroidism and a parathyroid adenoma complicating lithium treatment. The Lancet 2: 1129–1130.

42.

McKnight

Geddes

Goodwin

(2017) Short-and midterm side effects of lithium therapy. In: Malhi

Masson

Bellivier

(eds) The Science and Practice of Lithium Therapy. Cham: Springer, pp. 249–264.

43.

Malhi

Bassett

Boyce

et al . (2015) Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for mood disorders. Australian and New Zealand Journal of Psychiatry 49: 1087–1206.

44.

Malhi

Gershon

Outhred

(2016) Lithiumeter: Version 2.0. Bipolar Disorders 18: 631–641.

45.

Mata

Ramos

Bansal

et al . (2015) Prevalence of depression and depressive symptoms among resident physicians: A systematic review and meta-analysis. Journal of the American Medical Association 314: 2373–2383.

46.

Merry

(1977) Lithium and thyrotoxicosis. British Medical Journal 2: 765.

47.

Miller

Daniels

(2001) Association between lithium use and thyrotoxicosis caused by silent thyroiditis. Clinical Endocrinology 55: 501–508.

48.

Mizukami

Michigishi

Nonomura

et al . (1995) Histological features of the thyroid gland in a patient with lithium induced thyrotoxicosis. Journal of Clinical Pathology 48: 582–584.

49.

Moher

Liberati

Tetzlaff

et al . (2009) Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine 6: e1000097.

50.

Numata

Taniguchi

Harada

et al . (2002) Silent thyroiditis associated with short-term lithium therapy. General Hospital Psychiatry 24: 451–453.

51.

Oakley

Dawson

Whyte

(2000) Lithium: Thyroid effects and altered renal handling. Journal of Toxicology: Clinical Toxicology 38: 333–337.

52.

Onuigbo

MAC

Cuffy-Hallam

Wiley

et al . (2000) ‘Painless thyroiditis’: The rare phenomenon of lithium-associated thyrotoxicosis. Hospital Physician 36: 48–50.

53.

Pallisgaard

Frederiksen

(1978) Thyrotoxicosis in a patient treated with lithium carbonate for mental disease. Acta Medica Scandinavica 204: 141–143.

54.

Rabin

Evans

(1981) Exophthalmos and elevated thyroxine levels in association with lithium therapy. Journal of Clinical Psychiatry 42: 398–400.

55.

Rapaport

(1994) The effects of prolonged lithium exposure on the immune system of normal control subjects: Serial serum soluble interleukin-2 receptor and antothyroid antibody measurements. Biological Psychiatry 35: 761–766.

56.

Reus

Gold

Post

(1979) Lithium-induced thyrotoxicosis. American Journal of Psychiatry 136: 724–725.

57.

Sato

Taki

Honda

et al . (2013) Lithium toxicity precipitated by thyrotoxicosis due to silent thyroiditis: Cardiac arrest, quadriplegia, and coma. Thyroid 23: 766–770.

58.

Shimzu

Hirokawa

Manabe

et al . (1997) Lithium associated autoimmune thyroiditis. Journal of Clinical Pathology 50: 172–174.

59.

Shine

McKnight

Leaver

et al . (2015) Long-term effects of lithium on renal, thyroid, and parathyroid function: A retrospective analysis of laboratory data. The Lancet 386: 461–468.

60.

Sinnott

McIntyre

Pond

(1992) Granulomatous thyroiditis and lithium therapy. Australian and New Zealand Journal of Medicine 22: 84.

61.

Sirota

Cobin

Futterweit

et al . (1992) Hyperthyroidism in patients treated with lithium: Report of nine cases. Mount Sinai Journal of Medicine 59: 79–81.

62.

Siyam

Deshmukh

Garcia-Touza

(2013) Lithium-associated hyperthyroidism. Hospital Practice 41: 101–104.

63.

Szabadi

(1991) Thyroid dysfunction and affective illness. British Medical Journal 302: 923–924.

64.

Tan

Dhillon

Mohan

et al . (2013) Lithium-associated silent thyroiditis: Clinical implications. Australian and New Zealand Journal of Psychiatry 47: 965–966.

65.

Thompson

Baylis

(1986) Asymptomatic Graves’ disease during lithium therapy. Postgraduate Medical Journal 62: 295–296.

66.

Todd

Jerram

(1978) Thyrotoxicosis during lithium treatment. British Journal of Clinical Practice 32: 201–203.

67.

Valenta

Elias

Weber

(1981) Hyperthyroidism and propylthiouracil-induced agranulocytosis during chronic lithium carbonate therapy. American Journal of Psychiatry 138: 1605–1607.

68.

Vanderpump

Tunbridge

French

et al . (1995) The incidence of thyroid disorders in the community: A twenty-year follow-up of the Whickham Survey. Clinical Endocrinology 43: 55–68.

69.

Wilson

McKillop

Crocket

et al . (1991) The effect of lithium therapy on parameters thought to be involved in the development of autoimmune thyroid disease. Clinical Endocrinology 34: 357–361.

70.

Woolf

(1980) Transient painless thyroiditis with hyperthyroidism: A variant of lymphocytic thyroiditis? Endocrine Reviews 1: 411–420.

71.

Yamagishi

Yokoyama-ohta

(1999) A case of lithium-associated hyperthyroidism. Postgraduate Medical Journal 75: 188–189.

72.

Yassa

Saunders

Nastase

et al . (1988) Lithium-induced thyroid disorders: A prevalence study. Journal of Clinical Psychiatry 49: 14–16.

Lithium can cause hyperthyroidism as well as hypothyroidism: A systematic review of an under-recognised association

Abstract

Objective:

Methods:

Results:

Conclusion:

Keywords

Introduction

Methods

Search strategy

Inclusion criteria

Exclusion criteria

Study quality

Statistical analysis

Categorisation of study types

Results

Discussion

Clinical significance

Strengths and limitations

Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD’S

References