Introduction

Graves’ disease is the autoimmune thyroid disorder. Its onset is related to the interaction of genetic and environmental factors [1]. The primary clinical manifestation of the disease is Graves’ hyperthyroidism (GH) characterized by the anti-thyroid stimulating hormone (TSH) receptor antibodies (TRAb) in serum and overproduction of thyroxine (T4) and triiodothyronine (T3) in the gland [2].

The initial therapeutic approach of GH with the anti-thyroid drugs (ATD) aims to reduce the production of thyroid hormones (TH). A relatively high relapse rate of GH after ATD withdrawal has reported in over 50% of treated patients [3]. It has recommended that if the ATD treatment fails after 18 months, the radioiodine therapy (RIT) or thyroidectomy should be applied [3,4,5]. The purpose of RIT is to destroy sufficient thyroid tissue by β− particles emitted from radioiodine (131I). It is not possible to determine an appropriate 131I activity for achieving stable euthyroidism as the ideal outcome for GH. Therefore, hypothyroidism by lifetime thyroxine replacement therapy also represents a success of RIT [5]. The outcome of RIT mainly depends on the 131I uptake and the effective half-life (EHL) of 131I in the gland [6, 7]. The applied 131I activity, TH levels, TRAb, pretreatment with ATD, age, or sex may also affect the efficacy of RIT [8,9,10].

After an unexpected observation that lithium salts cause sedation in guinea pigs, lithium carbonate (LiCO3) was introduced in 1949 for stabilizing the mood in patients with bipolar psychoses [11]. It has been noticed that long-term treatment with LiCO3 in some patients induced goiter [12]. Studies have shown that LiCO3 inhibits the discharge rate of 131I from the thyroid [13] and blocks release of TH from the gland in both normal and hyperthyroid subjects [14]. It has also been shown that LiCO3 treatment for 14 days prolongs the thyroidal EHL of tracer doses of 131I [15]. Accordingly, the effect of LiCO3 on RIT outcome in GH was analyzed by several studies [16,17,18,19,20,21,22,23,24]. Some of them showed a higher cure rate in patients treated with 131I plus LiCO3 than in patients treated with 131I alone [16,17,18,19,20], while others contradicted the beneficial effect of LiCO3 on the overall RIT outcome [21,22,23,24]. Treatment with LiCO3 (6–21 days), its daily dose (750–900 mg), the beginning of the treatment before RIT (3–12 days), ATD withdrawal before RIT (3–5 days), and the applied 131I activities differed among the studies [16,17,18,19, 23]. The impact of LiCO3 on RIT outcome was mainly examined in patients with a recent onset of GH [16,17,18]. Concerning these studies, we chose 7-day treatment with LiCO3 (started 7 days after ATD withdrawal and 1 day before RIT) in a daily dose of 900 mg and applied 296–444 MBq of 131I-Na doses in order to improve efficacy of RIT.

The aim of this prospective study was to investigate the effect of short-term treatment with LiCO3 on the outcome of RIT in patients with long-lasting GH, its influence on the thyroid hormone levels 7 days after RIT, and possible side effects. The effect of age, duration of GH, TRAb, technetium-99m pertechnetate (99mTcO4 −) thyroid uptake (99mTcTU), and the applied 131I activities on the outcome of RIT in both groups were also studied.

Materials and methods

Study design

We prospectively included 60 patients with recurrent and long-lasting GH. Thirty patients were treated with 131I and LiCO3 (RI-Li group) while 30 patients with 131I alone (RI group). All patients were subjected to the same conditions of preparation to RIT and followed up for 12 months. The uncertain outcome of the study was approved by the Committee for postgraduate studies at the Faculty of Medicine Niš (reference number 04-835/12). The study was conducted from April 2012 to March 2016 at the Center of Nuclear Medicine, Clinical Center of Niš, Serbia. The patients were received detailed information about the procedure, possible outcome, and safety factors.

Eligibility criteria

The inclusion criteria were: both sexes aged 20–70 years, TRAb at the onset of GH, failure of ATD therapy, and absent or moderately present Graves’ ophthalmopathy (GO). We also included the patients with the gland size estimated by palpation as a grade 0 (normal-sized, invisible), grade 1 (slightly enlarged, visible), and grade 2 (moderately enlarged, highly visible). The exclusion criteria were: thyroid nodules suspicious for malignancy, severe GO, previous treatment with 131I, thyroidectomy, amiodarone or LiCO3, psychiatric or renal diseases.

Patients

The study included 49 females and 11 males aged 30–69 years (52.6 ± 9.05 year). The disease lasted longer than 2 years in 55 of 60 patients (range 7 months to 30 years). All patients were treated with ATD for 1–24 months to the remission of disease (8.62 ± 4.57 month). Values of 99mTcTU ranged from 0.50 to 21.5% (8.59 ± 5.12%).

Methods

After achieved remission, the ATD therapy was discontinued 7 days before RIT. The 99mTcTU test was performed 1 day before RIT (normal value 0.30–3.00%). We applied intravenously 74 MBq of 99mTcO4 − and after 20 min the thyroid uptake was measured by a Siemens e-cam gamma camera equipped with a low energy collimator. Thyroid uptake of 99mTcO4 − was calculated as the ratio of the thyroid counts and the net syringe counts that were both corrected for the background counts, radioactive decay of 99mTc, and attenuation of γ-photons in the air.

Treatment with LiCO3 started 1 day before RIT by administering orally 300 mg every 8 h. The treatment lasted for 7 days. Before 131I-NaI application, the serum level of lithium was measured (reference interval 0.300–1.30 mmol/l) (Flame Photometer IL945). Patients were obliged to report the side effects during treatment. Serum levels of sodium (Na), potassium (K), chlorine (Cl), bicarbonates (Bic), calcium (Ca), magnesium (Mg), urea (U), and creatinine (Cr) were determined 1 day before and at the end of LiCO3 treatment (Olympus AV680 analyzer, USA).

Serum levels of TSH, total T4 and T3 (TT4, TT3), free T4 and T3 (FT4, FT3), and TRAb were determined on the day of RIT. We repeated the measuring of TSH and TH 7 days after that (LKB-Wallac Clinic Gamma1272, Finland).

The dose of 131I-NaI was orally applied. It was determined concerning the gland size, i.e., 276, 370, and 444 MBq for grade 0, 1, and 2, respectively. Patients were followed up for 12 months. The serum levels of FT4, FT3, and TSH were measured after 1, 3, 6, 9, and 12 months. The success of RIT was defined as biochemically confirmed euthyroidism or hypothyroidism for 12 months and the failure as persistent hyperthyroidism.

Data analysis

The continuous variables are presented as mean ± standard deviation (SD) or mean ± standard error (SE) and median if required, and categorical variables as absolute numbers or percentage. Kolmogorov–Smirnov test is used to determine whether the values of continuous variables were normally distributed or not. Student’s t test for independent samples (normally distributed values) and Wilcoxon rank-sum test (non-normally distributed values) are used to assess the differences in patient’s baseline characteristics between groups. Chi-square test and Fisher’s exact test (if one cell in 2 × 2 table has the value less than 5) are used to compare categorical variables and therapeutic outcome among the groups. A repeated measure ANOVA is used to estimate the differences in mean values of variables measured before and after treatment within the groups. Logistic regression analysis is used to identify the independent variables that contribute to the prediction of the unfavorable modality of the dependent variable. Kaplan–Meier survival analysis is performed to estimate the cure time of disease and comparison between groups is assessed with the log-rank test. We considered the P value of <0.05 as significant. A statistical analysis is performed using the Statistical Package for the Social Sciences (SPSS Inc, Chicago IL, USA).

Results

The patient`s baseline characteristics of both groups on the day of RIT are shown in Table 1. They had similar age, duration of GH and ATD therapy, TRAb level and 99mTcTU before RIT, and received the same dose of 131I-NaI. The mean free and total TH as well as TSH serum levels did not differ between the groups (Table 1). However, TT4 and FT4 serum levels significantly increased while TSH level decreased 7 days after RIT in patients treated with 131I alone (Table 2). In contrast, there were no considerable changes in TT4, FT4, and TSH levels in RI-Li group, except decreased TT3 level (Table 2). Also, 7 days after RIT only mean FT4 serum level significantly differed among the groups (19.1 ± 6.93 vs. 12.7 ± 3.80 nmol/l, P < 0.001). The worsening of GO was not noticed for 12 months.

Table 1 Baseline characteristics of patients treated only with 131I and 131I with LiCO3 before radioiodine therapy
Full size table
Table 2 Thyroid hormones and TSH serum levels in patients treated only with 131I and 131I with LiCO3 on the day of radioiodine application and 7 days after that
Full size table

Lithium serum level was within the reference interval before RIT (0.571 ± 0.156 mmol/l; range 0.400–0.990 mmol/l). We did not notice the toxic side effects of lithium during the 7-day treatment. Comparison of electrolytes, urea, and creatinine levels 1 day before (all in normal ranges) and after the 7-day treatment with LiCO3 showed no significant changes (Table 3).

Table 3 Serum levels of electrolytes, urea, and creatinine before and after the 7-day-long treatment with LiCO3
Full size table

A higher percentage of cured patients was in RI-Li than in RI group (Fig. 1). The median cure time was shorter in the RI-Li group (1st month) than in RI group (3rd month) [mean ± SE 1.00 ± 0.00 month, CI (95%) 1.00–1.00 vs. 2.77 ± 0.426 month, CI (95%) 1.94–3.61; P = 0.000012]. Hypothyroidism was achieved faster in RI-Li (in five patients after 1 month) than in RI group (in eight patients after 3 months) but still, two of 22 patients in RI-Li group who were euthyroid after 1 month remained euthyroid 12 months after RIT. In contrast, all cured patients in RI group became hypothyroid after 6 months (Fig. 2).

Fig. 1
figure 1

Comparison of successful RIT outcome for 12 months among the patients’ groups treated only with 131I and 131I with LiCO3; *P < 0.0001, **P < 0.01

Full size image
Fig. 2
figure 2

Comparison of radioiodine therapeutic response for 12 months among the patients’ groups treated only with 131I and 131I with LiCO3; *P < 0.0001, **P < 0.01, ***P < 0.05

Full size image

Age, duration of GH, TRAb, and 99mTcTU (Table 4) did not have a predictive role for permanent hyperthyroidism 12 months after RIT in both groups. Despite that, we identified the dose of 131I-NaI as a predictor of RIT failure only in RI group (Table 4). Also, the applied 370 and 444 MBq activities of 131I were appropriate to successfully cure more patients in RI-Li than in RI group for 12 months (Table 5).

Table 4 Influence of the risk factors on a failure of radioiodine therapy after 12 months in patients treated only with 131I and 131I with LiCO3
Full size table
Table 5 Comparison between 131I-NaI dose and radioiodine therapy response for 12 months in patients treated only with 131I and 131I with LiCO3
Full size table

Discussion

Radioiodine therapy is effective, safe, and inexpensive for the treatment of GH [5]. The β− particles of 131I induce thyroid cell destruction in proportion to the quantity of radionuclide accumulated in the gland. The disease should transiently exacerbate after RIT because of the larger release of TH from the damaged gland [8]. Many factors may have the effect on RIT outcome in GH [6,7,8,9,10], but the most influential are 131I uptake and EHL of 131I in the gland [6, 7]. The reduced thyroid iodine pool, faster 131I turnover rate, and shorter EHL of 131I follow Graves’ hyperthyroidism. Hence, any attempt to prolong EHL of 131I in the gland may lead to improvement of RIT effectiveness.

Lithium carbonate is primarily used for the treatment of bipolar affective disorders, and its long-term administering may affect the thyroid function [25]. Dunkelmann et al. [26] showed that lithium enhanced the 131I half-life by 60.0%, and increased the applied thyroid radiation dose by 39.0% but did not affect the thyroid 131I uptake when was administered 885 mg/day for 2 weeks. These properties seem beneficial for RIT. Lithium carbonate was rarely utilized to improve the efficiency of RIT. It is not appropriate for the treatment of GH but only in the patients with very short EHL [26], for preventing transient exacerbation of the disease after ATD withdrawal before RIT [17] or in patients who do not tolerate ATD [27].

Presented results are in agreement with some results of our previous report obtained in a small number of patients with long-lasting GH [20]. We confirmed an increase of TT4 and FT4 serum levels and a decrease of TSH level 7 days after RIT only in patients treated with 131I alone. This phenomenon that occurs several days after RIT as a repercussion of gland irradiation was absent in patients treated with LiCO3. Other studies also reported this first protective effect of lithium whether it was given with [14, 16, 18] or without 131I [28,29,30]. Thus, in patients with recent onset of GH (≤6 months) treated with 131I plus LiCO3 (900 mg/day for 12 days), Bogazzi et al. [18] showed that FT4 serum level did not change significantly after RIT. They reported that FT4 level increased in patients treated with 131I alone reaching a peak between 3 and 5 days after RIT (P < 0.0001 for both times). Carlson et al. [29] found 26.0–45.0% inhibition of TT4 disappearance rate in hyperthyroid patients treated only with lithium for 6–7 days but without any change in TT4 disappearance slope in euthyroid individuals. In our study, mean TH serum levels and TSH level before and 7 days after RIT did not differ among the groups except the difference in mean FT4 level. Bogazzi et al. [18] also found a higher mean FT4 level in patients who received only 131I than in patients who received 131I and LiCO3. In the present study, mean TT3 level decreased significantly in RI-Li group 7 days after RIT which was not noticed in our previous report [20]. Similarly, Temple et al. [28] found that TT3 serum level fell somewhat more than TT4 level after lithium administering. So, they postulated that lithium might inhibit the conversion of thyroxine to triiodothyronine.

The use of LiCO3 is accompanied by controversies for the treatment of bipolar affective disorders because of life-long dose monitoring and the appearance of both acute and chronic side effects [25, 31]. Lithium carbonate has a simple pharmacokinetics, and after an oral administering reaches a peak of serum concentration for 4–12 h. It thoroughly distributes into interstitial fluid and ultimately slowly enters into the intracellular fluid [32]. Lithium exclusively excretes via the kidneys and any acute or chronic disturbance of the excretion increases its serum concentration and may become toxic [31, 32]. The pathophysiology of renal, parathyroid, and thyroid dysfunctions during its long-term use is unclear [33]. We showed no side toxic effects of lithium for 7 days of use. Mean serum levels of electrolytes, urea, and creatinine before and after 7-day treatment were similar. There are insufficient reports on LiCO3 side effects when its short-term use (400–900 mg/day) with 131I is combined. The studies of Turner et al. [15] and Bogazzi et al. [18] reported that the side effects of LiCO3 were not relevant. In other studies, only one of 54 patients had mild nausea [16] or insignificant appearance of nausea and neck sensitivity [17]. We noticed that LiCO3 supporting RIT had no impact on GO. To our knowledge, only one study reported that the patients had no worsening of GO for 12 months after RIT [16].

We showed the remarkable effect of LiCO3 on RIT outcome in patients with long-lasting GH. Lithium-treated patients were cured more rapidly (27/30 patients were euthyroid or hypothyroid after 1 month) than those treated only with 131I-NaI (12/30 patients were euthyroid after one and 22 patients were hypothyroid after 6 months). Although the daily dose of LiCO3, duration of treatment and its beginning before RIT, and the applied 131I activities [16,17,18,19,20,21,22,23,24] differed among studies, our results are in agreement with the results of some other studies [16,17,18,19]. The cited studies [16,17,18], except the study of Martin et al. [19], included patients with recent onset of GH. After 5 days of ATD withdrawal, Bogazzi et al. [16] treated a newly diagnosed GH patients with LiCO3 for 6 days (900 mg/day) starting on the day of 131I application (521 ± 148 and 556 ± 141 MBq in RI and RI-Li group, respectively). After 1 year, the success of RIT was 72.0% in RI and 83.0% in RI-Li group (P = 0.02). Martin et al. [19] showed the success in 93.1% of patients treated with 131I-NaI (500 MBq) and LiCO3 for 10 days (800 mg/day, starting 3 days before RIT), as well in 83.5% of patients treated only with 131I-NaI (P = 0.049). This improvement of RIT efficiency occurs due to the extension of EHL of 131I in the gland by LiCO3 [13,14,15, 26] and the increment of radiation dose in the thyroid [18]. However, others showed that LiCO3 did not affect the outcome of RIT [21,22,23,24]. Two studies [21, 22] explained such result by a high cure rate in both groups and a small number of treated patients. Also, Bal et al. [23] reported that a success of RIT was 71.6% in RI-Li and 70.5% in control group after 12 months (P > 0.05). In that study, ATD withdrawal was 3–4 days before RIT, the treatment with LiCO3 started on the day of RIT (900 mg/day for 21 days), and the low doses of 131I-NaI in both groups (196 ± 55.5 and 211 ± 59.2 MBq) were applied. Comparison of the designs and results of the cited studies show that the insignificant effect of LiCO3 on RIT outcome occurs when a low dose of 131I-NaI was applied [23], or when the treatment with LiCO3 started on the day of RIT [16, 23]. In the second case, the effect of LiCO3 on the follicular cells level is diminished probably due to it slowly entering the intracellular fluid [32], and its incomplete accumulation in the tissue before 131I was applied.

The important finding of our study is that the number of euthyroid patients slowly decreases in the lithium-treated group, but not all cured patients become hypothyroid for 12 months. On the contrary, the number of euthyroid patients who did not receive LiCO3 rapidly declines and all of them become hypothyroid for 6 months. That may suggest a protective effect of LiCO3 on preventing early induction of hypothyroidism.

Age, duration of GH, TRAb, and 99mTcTU were not determined as predictors of permanent GH in both groups. It is not possible to establish a direct connection between the risk factors and genetic susceptibility because of presence and interaction of several of these factors in a patient before the onset of disease [1]. Despite that, the dose of 131I-NaI had a predictive role in RIT failure only in RI group. Finally, comparison of the RIT response and the applied 131I activities points that the short-term treatment with LiCO3 provides the success of RIT in more patients with the low doses of 131I-NaI. This finding could confirm that LiCO3 increases radiation dose in the gland and leads to a higher level of tissue irradiation. To our knowledge, only Martin et al. [19] showed that the 131I activity, gender, and age correlated with a high probability of successful treatment of GH with 131I and LiCO3, although, they applied the high doses of 131I-NaI in both groups (510–597 and 537–595 MBq).

Conclusion

Although the patients with GH should not be treated with ATD more than 18 months, it is often disregarded in clinical practice. Generally, the long-lasting GH and comorbidities make it difficult to apply the radioiodine or surgical therapy for the finally cure of the disease. Lithium carbonate as an adjunct to 131I is rarely used in daily practice except for preventing transient worsening of the disease after RIT and mainly in patients with newly diagnosed GH. Despite that, our perspective study included the patients with long-lasting GH who were treated with relatively low doses of 131I-NaI. The study indicates that the short course of LiCO3 treatment in combination with 131I has the beneficial effect on RIT outcome in these patients. A successful outcome is achieved faster in lithium-treated patients than in those treated with 131I alone. Lithium carbonate prevents transient worsening of hyperthyroidism after 131I application. The treatment with LiCO3 for 7 days is safe in patients with normal renal function and serum electrolytes levels. Also, age, duration of GH, TRAb, and 99mTcTU have no influence on RIT outcome in patients treated and not treated with LiCO3, but the dose of 131I-Na has a predictive role for a failure of RIT in patients treated only with 131I.