Background
The relationship between affective disorders, mood-stabilisers and thyroid dysfunction is complex and poorly understood. Symptoms of thyroid dysfunction can overlap with symptoms of affective disorder, destabilise mood, and impact physical health. Subjective symptoms and biochemical abnormalities may not always match, especially when changes in thyroid function are only mild. Therefore, diagnosis and treatment of both hypothyroidism and hyperthyroidism in individuals with affective disorders remain complex. For lithium, a first-line treatment for bipolar disorder, an impact on thyroid function was first described in 1968. Since that time, it has become evident that lithium is much more frequently associated with hypothyroidism than hyperthyroidism. But even for lithium, many aspects of how associated thyroid dysfunction should be handled remain unclear.
Aims
The overall aim of this thesis was, in five studies, to examine aspects of the diagnosis and treatment of thyroid dysfunction in individuals with affective disorders, with a particular focus on lithium. The individual aims of the five studies were to
- determine if lithium-associated hypothyroidism was reversible in individuals who had discontinued lithium.
- identify patterns and trends in thyroid hormone replacement therapy prescribed for individuals with bipolar or schizoaffective disorder.
- assess whether elevated thyroxine concentrations (hyperthyroxinaemia) were a risk factor for lithium intoxication caused by a change in tubular renal function.
- examine the incidence rate and aetiology of lithium-associated hyperthyroidism in individuals with bipolar or schizoaffective disorder.
- explore the attitudes of practising clinicians towards the diagnosis and treatment of subclinical hypothyroidism in individuals with or without affective disorder or anxiety.
Methods
Studies 1–4 were part of the LiSIE (Lithium - Study into Effects and Side Effects) retrospective cohort study. LiSIE compares the effects and adverse effects of lithium treatment and other mood stabilisers in the Norrbotten Region and the Region of Västerbotten over a time period of up to 21 years between 1997–2017. For our studies, we used data from the Norrbotten Region only. Study 5 used a three-round modified Delphi consensus-building process. Study 5 was conducted with clinicians from three specialties, general practice, endocrinology and psychiatry, from two countries with similar health care systems, Sweden and the UK.
Results
Study 1: Of 1340 potentially eligible individuals with lithium treatment, 90 individuals (who had developed hypothyroidism while treated with lithium and later discontinued lithium), were included. Of these, 27% had overt hypothyroidism at the start of thyroid hormone replacement therapy. Of the 85 individuals available for follow-up, 41% stopped thyroid hormone replacement therapy after lithium discontinuation. Only six individuals reinstated thyroid hormone replacement therapy subsequently. Only one had overt hypothyroidism.
Study 2: Of 1564 potentially eligible individuals with bipolar or schizoaffective disorder, 291 (27%) had received thyroid hormone replacement therapy at some point during the 21-year review period. In 41% of cases, thyroid hormone replacement therapy was started for subclinical hypothyroidism. At the start of thyroid hormone replacement therapy, the median thyroid stimulating hormone (TSH) concentration was 6.0 (IQR 4.0) mIU/L. The median free serum thyroxine (fT4) was 11.8 (IQR 3.9) pmol/L. The median TSH concentration at the start of thyroid hormone replacement therapy decreased annually by 0.10 mIU/L, being significantly higher in individuals treated with lithium than in individuals treated with other mood stabilisers.
Study 3: Of 1562 potentially eligible individuals with bipolar or schizoaffective disorder, 53 individuals had experienced a total of 65 episodes of unintentional lithium intoxication during the review period. In nine episodes, there was elevated fT4 at the time of lithium intoxication, corresponding to an incidence of 1.3 episodes/1000 person-years. For all nine episodes of unintentional lithium intoxication, we could identify alternative explanations that were more plausible than hyperthyroxinaemia.
Study 4: In 1562 individuals with bipolar disorder or schizoaffective disorder, we identified 16 episodes of hyperthyroidism, corresponding to an incidence rate of 0.9 episodes/1000 person-years. Individuals who had concurrently been exposed to lithium, had an incidence rate of 1.3 episodes/1000 person-years. Individuals who had been previously exposed to lithium had an incidence rate of 0.8/1000 person-years. Individuals who had never been exposed to lithium (lithium naïve) had a 0.5/1000 person-years incidence rate. There were no significant differences in the risk ratios for individuals with concurrent or previous exposure compared to lithium-naïve individuals, neither for hyperthyroidism overall, nor for thyrotoxicosis or thyroiditis.
Study 5: For the expert panel, 60 clinicians; 20 general practitioners, 20 endocrinologists and 20 psychiatrists were recruited. Fifty-three (88%) participants completed all three rounds. The participants reached a consensus on five of the 26 practice statements. The participants agreed that (a) repeated testing was required for the diagnosis of subclinical hypothyroidism, (b) antibody screening should usually occur, and (c and d) antibody screening would strengthen the indication for thyroid hormone replacement therapy in both individuals with and without affective disorder or anxiety. The participants disagreed with (e) requiring a TSH threshold of ≥ 20 mIU/L before starting thyroid hormone replacement therapy.
Conclusions
Study 1: In most cases, lithium-associated hypothyroidism appears reversible. Therefore, thyroid hormone replacement therapy could be discontinued more often once lithium is stopped.
Study 2: In most cases, thyroid hormone replacement therapy was started with mild or absent thyroid function changes. The TSH level at which thyroid hormone replacement therapy was initiated decreased over time. When starting thyroid hormone replacement therapy for subclinical hypothyroidism in people with bipolar or schizoaffective disorder, clinicians must carefully weigh the benefits and risks.
Study 3: Lithium intoxication with simultaneously elevated fT4 is uncommon. A direct causal link between elevated fT4 and altered tubular renal function remains elusive. An increased frequency of routine thyroid function tests is unlikely to decrease the risk of lithium intoxication.
Study 4: Lithium-associated hyperthyroidism is uncommon. The risk of hyperthyroidism does not differ significantly between lithium-exposed and lithium-naïve individuals.
Study 5: Attitudes toward diagnosing and treating subclinical hypothyroidism remain diverse. A threshold of an TSH of at least 20 mIU/L for thyroid hormone replacement therapy start, suggested in a previously published guideline, was deemed too high. As the evidence regarding diagnosis and treatment of subclinical hypothyroidism remains limited, future guidelines should consider the views of a broad range of practising clinicians to increase their clinical acceptability and usefulness.