Baseline Thyrotrophin and Subclinical Hyperthyroidism
We identified 414 potential patients with endogenous subclinical hyperthyroidism from an electronic search of clinic records, Clinical Biochemistry laboratory database and records of radioactive iodine uptake and thyroid ultrasound scan reports in our institution between January 2003 and December 2010. After a review of individual case notes, we excluded patients who were receiving treatment with antithyroid drugs (n = 36), levothyroxine (n = 23), amiodarone (n = 5) or patients who had undergone treatment with radioactive iodine within 12 months of diagnosis (n = 2). We also excluded patients with a diagnosis of hypopituitarism (n = 2), women who were diagnosed during pregnancy (n = 3), patients with transient subclinical hyperthyroidism not confirmed on repeat testing (n = 11) and patients with untraceable hospital records (n = 9). Thus, a final number of 323 patients were included in the study.
The clinical characteristics of patients at baseline are presented in Table 1. Of the 323 patients, 63·5% had a TSH 0·10–0·39 mU/l (grade I), while 36·5% had TSH < 0·10 mU/l (grade II). Disagreement in TSH grade between the initial two baseline tests was seen in 14 patients and was resolved with the result of a third baseline test where available (n = 9) or in favour of the first of the two baseline tests if a third test was unavailable (n = 5). There was no difference in age distribution or gender between the two TSH groups. The majority of patients (69·3%) had toxic nodules, comprising toxic multinodular goitre (67·5%) and solitary toxic nodules (1·8%). Graves' disease accounted for just 5·6% of patients and was significantly more common in patients with grade II than grade I subclinical hyperthyroidism. The aetiological diagnosis was indeterminate in 81 cases despite clinical, laboratory and scintigraphic evaluation. FT3 measurements were unavailable at the first baseline evaluation in 22 patients but later confirmed to be normal on a subsequent measurement thereby excluding T3 toxicosis in these patients.
The duration of follow-up ranged from 6 to 93 months with a mean follow-up duration of 32 months. Only 38 patients (11·8%) progressed to overt hyperthyroidism, while the majority of patients either reverted to normal thyroid status (31·6%) or had persistent subclinical hyperthyroidism (56·7%). Eight of the patients with persistent subclinical hyperthyroidism, including seven with grade II and one with grade I subclinical hyperthyroidism, were offered treatment because of the development of atrial fibrillation (n = 2), congestive cardiac failure (n = 2), osteoporosis (n = 1) and suggestive symptoms of hyperthyroidism (n = 3). Amongst those patients who developed hyperthyroidism, 25 patients (65·8%) had elevated FT4 and FT3, 4 patients (10·5%) had a raised FT4 with normal FT3, while nine patients (23·7%) had elevated FT3 with normal FT4, that is, T3 toxicosis. The annual progression rate to overt hyperthyroidism was 0·6–3·7% with cumulative progression rates of 2·2% at 1 year, 5·9% in the second year, 9% in the third year and 11·1% at the end of the fifth year. Table 2 shows the relationship between outcomes and baseline TSH status. The proportion of patients who progressed to overt hyperthyroidism was significantly higher amongst patients with grade II (TSH < 0·10 mU/l) than those with grade I (TSH 0·10–0·39 mU/l) subclinical hyperthyroidism (20·3% vs 6·8%, P < 0·001, Chi square test).
Kaplan–Meier curves showed significantly faster rates of progression of overt hyperthyroidism in patients with grade II than grade I hyperthyroidism (P < 0·001, log rank test, Fig. 1).
(Enlarge Image)
Figure 1.
Kaplan–Meier curves of the incidence of overt hyperthyroidism according to the grades of baseline TSH levels.
We performed stepwise Cox proportional univariate and multivariate analysis to determine the association of various clinical and biochemical factors with progression to overt hyperthyroidism (Table 3). In the univariate analysis, only a baseline TSH < 0·10 mU/l was significantly associated with progression to overt hyperthyroidism. Factors that were included in the adjusted Cox model included baseline TSH (<0·1 mU/l or 0·1–0·39 mU/l), age (<65 years or ≥ 65 years), gender (male or female), aetiological diagnosis (Graves' disease, toxic nodular goitre or indeterminate diagnosis) and baseline FT4 grouped according to tertiles. Of these factors, only a baseline TSH < 0·10 mU/l remained significantly associated with the development of overt hyperthyroidism (hazard ratio 3·4, confidence interval 1·6–7·0).
Results
Patient Characteristics
We identified 414 potential patients with endogenous subclinical hyperthyroidism from an electronic search of clinic records, Clinical Biochemistry laboratory database and records of radioactive iodine uptake and thyroid ultrasound scan reports in our institution between January 2003 and December 2010. After a review of individual case notes, we excluded patients who were receiving treatment with antithyroid drugs (n = 36), levothyroxine (n = 23), amiodarone (n = 5) or patients who had undergone treatment with radioactive iodine within 12 months of diagnosis (n = 2). We also excluded patients with a diagnosis of hypopituitarism (n = 2), women who were diagnosed during pregnancy (n = 3), patients with transient subclinical hyperthyroidism not confirmed on repeat testing (n = 11) and patients with untraceable hospital records (n = 9). Thus, a final number of 323 patients were included in the study.
The clinical characteristics of patients at baseline are presented in Table 1. Of the 323 patients, 63·5% had a TSH 0·10–0·39 mU/l (grade I), while 36·5% had TSH < 0·10 mU/l (grade II). Disagreement in TSH grade between the initial two baseline tests was seen in 14 patients and was resolved with the result of a third baseline test where available (n = 9) or in favour of the first of the two baseline tests if a third test was unavailable (n = 5). There was no difference in age distribution or gender between the two TSH groups. The majority of patients (69·3%) had toxic nodules, comprising toxic multinodular goitre (67·5%) and solitary toxic nodules (1·8%). Graves' disease accounted for just 5·6% of patients and was significantly more common in patients with grade II than grade I subclinical hyperthyroidism. The aetiological diagnosis was indeterminate in 81 cases despite clinical, laboratory and scintigraphic evaluation. FT3 measurements were unavailable at the first baseline evaluation in 22 patients but later confirmed to be normal on a subsequent measurement thereby excluding T3 toxicosis in these patients.
Progression to Overt Hyperthyroidism
The duration of follow-up ranged from 6 to 93 months with a mean follow-up duration of 32 months. Only 38 patients (11·8%) progressed to overt hyperthyroidism, while the majority of patients either reverted to normal thyroid status (31·6%) or had persistent subclinical hyperthyroidism (56·7%). Eight of the patients with persistent subclinical hyperthyroidism, including seven with grade II and one with grade I subclinical hyperthyroidism, were offered treatment because of the development of atrial fibrillation (n = 2), congestive cardiac failure (n = 2), osteoporosis (n = 1) and suggestive symptoms of hyperthyroidism (n = 3). Amongst those patients who developed hyperthyroidism, 25 patients (65·8%) had elevated FT4 and FT3, 4 patients (10·5%) had a raised FT4 with normal FT3, while nine patients (23·7%) had elevated FT3 with normal FT4, that is, T3 toxicosis. The annual progression rate to overt hyperthyroidism was 0·6–3·7% with cumulative progression rates of 2·2% at 1 year, 5·9% in the second year, 9% in the third year and 11·1% at the end of the fifth year. Table 2 shows the relationship between outcomes and baseline TSH status. The proportion of patients who progressed to overt hyperthyroidism was significantly higher amongst patients with grade II (TSH < 0·10 mU/l) than those with grade I (TSH 0·10–0·39 mU/l) subclinical hyperthyroidism (20·3% vs 6·8%, P < 0·001, Chi square test).
Survival Analysis
Kaplan–Meier curves showed significantly faster rates of progression of overt hyperthyroidism in patients with grade II than grade I hyperthyroidism (P < 0·001, log rank test, Fig. 1).
(Enlarge Image)
Figure 1.
Kaplan–Meier curves of the incidence of overt hyperthyroidism according to the grades of baseline TSH levels.
Univariate and Multivariate Analysis
We performed stepwise Cox proportional univariate and multivariate analysis to determine the association of various clinical and biochemical factors with progression to overt hyperthyroidism (Table 3). In the univariate analysis, only a baseline TSH < 0·10 mU/l was significantly associated with progression to overt hyperthyroidism. Factors that were included in the adjusted Cox model included baseline TSH (<0·1 mU/l or 0·1–0·39 mU/l), age (<65 years or ≥ 65 years), gender (male or female), aetiological diagnosis (Graves' disease, toxic nodular goitre or indeterminate diagnosis) and baseline FT4 grouped according to tertiles. Of these factors, only a baseline TSH < 0·10 mU/l remained significantly associated with the development of overt hyperthyroidism (hazard ratio 3·4, confidence interval 1·6–7·0).
Source...