Neuropsychological Function and Symptoms in ... - Oxford Journals

0021-972X/06/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 91(1):145–153 Copyright © 2006 by The Endocrine Society doi: 10.1210/jc.2005-1775

Neuropsychological Function and Symptoms in Subjects with Subclinical Hypothyroidism and the Effect of Thyroxine Treatment Rolf Jorde, Knut Waterloo, Hilde Storhaug, Audhild Nyrnes, Johan Sundsfjord, and Trond Geir Jenssen Institute of Clinical Medicine (R.J., K.W.), University of Tromsø, 9037 Tromsø, Norway; Departments of Internal Medicine (H.S., A.N.) and Clinical Chemistry (J.S.), University Hospital of North Norway, 9038 Tromsø, Norway; and Department of Nephrology (T.G.J.), National Hospital, 0027 Oslo, Norway Results: The mean ⫾ SD serum TSH in the SHT and control group were 5.57 ⫾ 1.68 and 1.79 ⫾ 0.69 mIU/liter, respectively. There were no significant differences in cognitive function and hypothyroid symptoms between the two groups, but those with SHT scored significantly better than the controls on the GHQ-30. At the end of the intervention study, serum TSH in the T4 group (n ⫽ 36) and the placebo group (n ⫽ 33) were 1.52 ⫾ 1.51 and 5.42 ⫾ 1.96 mIU/liter, respectively. T4 substitution had no effect on any of the parameters measured.

Objective: Our objective was to examine the relation between neuropsychological function and subclinical hypothyroidism (SHT), defined as serum TSH of 3.5–10.0 mIU/liter and normal serum free T4 and free T3 levels, and to study the effect of T4 supplementation. Subjects: A total of 89 subjects (45 males) with SHT and 154 control subjects (72 males) were recruited from a general health survey (the fifth Tromsø study). Sixty-nine of those with SHT were included in a placebo-controlled, double-blind intervention study with T4 medication for 1 yr.

Conclusion: In subjects with SHT where the serum TSH level is in the 3.5–10.0 mIU/liter range, there is no neuropsychological dysfunction, and compared with healthy controls, there is no difference in symptoms related to hypothyroidism. (J Clin Endocrinol Metab 91: 145–153, 2006)

Main Outcome Measures: We used fourteen tests of cognitive function, Beck Depression Inventory, General Health Questionnaire, and a questionnaire on hypothyroid symptoms.

I

N PATIENTS WITH classical symptoms of hypothyroidism and with elevated level of serum TSH combined with a low serum free T4, the decision to start treatment with T4 is usually easy. However, whether to treat subjects with subclinical hypothyroidism (SHT), defined as an elevated serum TSH level with serum free T4 and free T3 levels within the normal range and no overt symptoms of hypothyroidism, is controversial (1– 4) and has been the topic of several recent review articles (5– 8). The effect of T4 supplementation in SHT has been most thoroughly studied regarding the lipid profile, and there appears to be a slight effect on serum total and low-density lipoprotein cholesterol levels (9). However, studies regarding effects of T4 supplementation on clinical symptoms and neuropsychological dysfunction in SHT are less conclusive with reports showing a clear positive effect (10, 11), a moderate or no effect (12), and even an adverse effect (13). By definition, subjects with SHT should not have overt symptoms of hypothyroidism. However, because the disease usually develops gradually, the symptoms may go unrecognized by the patients and their families. Symptoms compatible with hypothyroidism are also seen in subjects with

normal thyroid function (14), and accordingly, the presence of such symptoms in subjects with SHT may have causes other than thyroid dysfunction. Many patients are therefore diagnosed with SHT because thyroid function tests are often performed in subjects with nonspecific symptoms such as fatigue and depression. Therefore, to assess the true association between reduced thyroid function and symptoms, which may be of importance in the discussion on treatment with T4, one should not include subjects recruited from clinical practice. In addition, the subjects should not be aware of their thyroid status before testing. We recently had the opportunity to study such a group. Thus, the Tromsø study, which is an epidemiological health survey, was performed for the fifth time in 2001, and serum TSH was measured in 7954 subjects. From this cohort, subjects with SHT and normal controls were recruited and examined with a broad range of neuropsychological tests and symptom scores. To study a group of subjects where the thyroid function was only moderately reduced, we included subjects with serum TSH in the range of 3.5–10.0 mIU/liter. Subjects and Methods The fifth Tromsø study was performed as a general health survey in 2001 in a manner similar to the previous ones (15). All men and women older than 29 yr, living in the municipality of Tromsø and that participated in the second phase of the fourth Tromsø study (16) or became 30, 40, 45, 60, or 75 yr old during 2001, were invited to participate. All subjects filled out a health questionnaire, and they were also asked whether they would prefer not to be invited to additional studies based on results from the present one. Nonfasting blood samples were drawn and analyzed for serum TSH.

First Published Online November 1, 2005 Abbreviations: BDI, Beck Depression Inventory; BMI, body mass index; CalCAP, California Computerized Assessment Package; GHQ, General Health Questionnaire; SHT, subclinical hypothyroidism; WAIS, Wechsler Adult Intelligence Scale. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

145

146

J Clin Endocrinol Metab, January 2006, 91(1):145–153

Subjects with serum TSH level between 3.5 and 10.0 mIU/liter were invited to a follow-up examination at the Clinical Research Unit at the University Hospital of Tromsø. Those who reported a history of coronary infarction, angina pectoris, or stroke in the questionnaire, those participating in other follow-up studies, those using thyroid medication, and those above the age of 80 were not invited. The hospital records were also reviewed to exclude subjects with serious diseases not reported in the questionnaire. For each subject with serum TSH between 3.5 and 10.0 mIU/liter, a randomly selected age- and sex-matched control subject from the fifth Tromsø study with serum TSH in the range 0.50 –3.49 mIU/liter was also invited. The invitation letter informed about the purpose of the study but did not disclose the subject’s TSH status. At the follow-up visit, blood was drawn in the nonfasting state and serum samples analyzed for TSH, free T4, and free T3. A clinical examination was performed, height and weight were measured in light clothing wearing no shoes, and body mass index (BMI) was calculated as weight (kg) divided by squared height (m2). A questionnaire focusing on hypothyroid symptoms was administered. Those who still had serum TSH between 3.5 and 10.0 mIU/liter and with free T4 and free T3 within the reference range, and without obvious clinical symptoms of hypothyroidism, were considered to have SHT. They were informed that their thyroid status was similar to that at the Tromsø study and invited to additional examinations. The control subjects who at the follow-up visit still had serum TSH within the 0.50 –3.49 mIU/liter range and with normal serum free T4 and free T3 were similarly informed that their thyroid status was unchanged and invited to additional examinations. Those consenting were, on a separate day, in the nonfasting state, examined with neuropsychological tests for cognitive and emotional function. The tests were administered in two sessions on the same day, each lasting 1 or 2 h and separated by a coffee break. The examiners, who were nurses extensively trained and certified by a clinical neuropsychologist (K.W.) and an experienced technician, carried out the tests in a standardized fashion and in the same order. The examiners were blinded to the TSH status of the participants. The subjects were then informed about their thyroid status. Those with SHT were invited to a 12-month intervention study. Those willing were randomized to placebo or T4. The T4 tablets were in 25-, 50-, and 100-␮g doses, and the corresponding placebo tablets were identical looking. Each subject was to take three tablets every day, and by combining T4 and placebo tablets, daily doses of T4 could be 25, 50, 75, 100, 125, 150, or 175 ␮g. During the first 6 wk, all subjects in the T4 group were given 50 ␮g daily, and for the following 6 wk 100 ␮g daily. Thereafter, the T4 dose was given according to the TSH levels, aiming at a level between 0.5 and 1.5 mIU/liter using the following algorithm: serum TSH less than 0.50 mIU/liter and free T4 more than 29 pmol/liter, reduce T4 dose by 50 ␮g; serum TSH less than 0.50 mIU/liter and free T4 less than 30 pmol/liter: reduce T4 dose by 25 ␮g; serum TSH 0.50 –1.50 mIU/liter, unchanged T4 dose; serum TSH 1.51– 4.00 mIU/liter, increase T4 dose by 25 ␮g; serum TSH more than 4.00 mIU/liter, increase T4 dose by 50 ␮g (if good compliance). Nonfasting blood samples for serum TSH, free T4, and free T3 were drawn after 3, 6, 9, and 12 months. On the blood sampling days, the T4 or placebo tablets were taken after the blood sampling. The TSH results were available after 1 or 2 d, and T4 (or placebo) tablets for the next 3-month period were sent by mail. The subjects continued their previous dose until they received the new tablets. At each visit, unused tablets from the previous period were counted. The compliance rate was calculated as number of tablets not returned divided by the corresponding number of days. After 12 months, a questionnaire focusing on change in symptoms since the first visit was administered and the neuropsychological tests repeated.

Tests of cognitive function For attention, sustained attention, and working memory, we used the Digit Span forward and backward test (subtests from Wechsler Memory Scale-Revised) (17) and the Seashore Rhythm test from the HalsteadReitan test battery (18). For psychomotor/cognitive speed, we used the Trail Making test, part A (18); the Stroop Color-Word test, parts 1 and 2 (reading speed) (scores for 1 and 2 added together) (19), modified version (20); and the Digit Symbol test (21). For memory, we used the verbal and visual paired associates imme-

Jorde et al. • Neuropsychological Function and TSH

diate and 30-min delayed recall (from Wechsler Memory Scale-Revised) (17) and verbal recall test and a word list consisting of 12 words, a subtest from California Verbal Learning Test (22). For language/word fluency, we used the Controlled Oral Word Association test with words beginning with the letters F, A, and S (23). For cognitive flexibility/executive function, we used the Trail Making test, part B (18), and the Stroop Color-Word test, part 3 (color-word interference effect) (19, 20). For speed of information processing, we used the California Computerized Assessment Package (CalCAP) (24). For intelligence, we used the subtest Vocabulary from the Wechsler Adult Intelligence Scale (WAIS) (21). A composite score for cognitive function was made by adding together the Z-scores for the following seven tests: Digit Span forward, Digit Span backward, Stroop test parts 1 and 2 (scores added together), Digit Symbol test, verbal recall, visual recall, and Stroop test part 3. If a negative score was favorable, the Z-score was multiplied by ⫺1.

Tests of emotional function Depressed mood was measured with the Beck Depression Inventory (BDI), which is a self-completed questionnaire of 21 items in multiplechoice format (25). The items constituting the BDI have been divided into two subscales. The first, Cognitive-Affective, assesses the mental aspect of depression (items 1–13). The second, Somatic-Vegetative, measures vegetative and somatic symptoms (items 14 –21). For the BDI, the total score and the subscale scores were obtained for each patient. Mental health status (or psychological distress) was assessed by the General Health Questionnaire (GHQ), which is a generic health instrument and may thus be applied across different diseases or conditions (26). The GHQ comes in several versions, and we applied the GHQ-30 version. On each of the 30 items, subjects are asked to compare their perceived state of health with four standard answers in the questionnaire. If scored according to the GHQ-scoring method (0 – 0 – 1 – 1), it may be used as a screening instrument identifying cases with nonpsychotic psychiatric disturbances or as a measure of psychiatric disturbance in the population. Because the objective of this study was not to identify cases but to obtain quantitative measures of well-being and mental health, we applied the Likert-scoring method (0 – 1 – 2 – 3), which also allows factor-scoring within the five subscales of GHQ-30. These factors are identified as factor A (anxiety), factor B (feelings of incompetence), factor C (depression, hopelessness), factor D (difficulty in coping), and factor E (social dysfunction) (27).

Symptom score A questionnaire containing 19 questions related to hypothyroid symptoms was administered at inclusion. A total symptom score was created by adding together the number of present symptoms (28). At the end of the study, a similar questionnaire with 10 questions on change of symptoms was administered, and a change score was calculated, giving 1 point if the change was in the better (more euthyroid) direction, ⫺1 point if the change was in the hypothyroid direction, and 0 point if there was no change.

Laboratory analyses Serum TSH levels from the fifth Tromsø study and the study on the SHT and control groups were analyzed with the Modular E instrument (Hoffmann-La Roche, Basel, Switzerland) with reference range 0.2– 4.0 mIU/liter. Accordingly, the values from this assay were used when including the subjects. At the start of the intervention study, our Department of Clinical Chemistry changed to the AxSYM (Abbott, IL) instrument with inherent reagents. Because we knew about this change in advance, serum samples from the SHT/control group study were stored and later analyzed with the new assay. All serum TSH, free T4, and free T3 values presented therefore represent analysis on the AxSYM instrument with reference values 0.20 – 4.20 mIU/liter, 9 –22 pmol/liter, and 2.8 –7.1 pmol/liter, respectively.

Statistical analyses Normal distribution was evaluated with determination of skewness and kurtosis and visual inspection of histograms. The scores for tests of



cognitive function were considered normally distributed except for Seashore Rhythm test, Trail Making B, Stroop Color-Word test parts 1 and 2, and the Stroop Color-Word test part 3. After logarithmic transformation, these latter variables assumed normal distribution and were applied as such when used as dependent variables. To test for interactions, factor analyses with scores for cognitive function as dependent variables; TSH group (SHT or control), gender, smoking status (current smoker/nonsmoker) as factors; and age and BMI as covariables were performed. This revealed significant effects of smoking status as well as interactions between smoking status and TSH group regarding several of the tests for cognitive function, and additional analyses regarding cognitive function were therefore done in nonsmokers only because there were too few smokers to allow significant relations to be found in that group. None of the tests for emotional function were normally distributed, nor did they assume normal distribution after logarithmic transformation. When compared with the Mann-Whitney U test in the control and the SHT groups separately, smokers and nonsmokers did not differ significantly on any of the tests for emotional function [P ⫽ 0.07 for GHQ-E (social dysfunction) in the SHT group; all other P values ⬎ 0.33]. Similarly, the symptom scores were not normally distributed, and there were no significant effects of smoking status when tested as above. Smokers and nonsmokers were therefore evaluated together regarding emotional function and symptom scores, but the results are also given for the nonsmokers separately. Cognitive function was compared between the two groups with Student’s t test and also with a general linear model with the parameter in question as dependent variable and with TSH group (SHT or control) and gender as factors and age and BMI as independent variables. A multiple linear regression model was used to assess independent predictors of the test scores for cognitive function. Because of the high correlation between free T4 and free T3, two models were used, with gender, age, BMI, serum TSH, and serum free T4 (or serum free T3) as covariables. For emotional function and the symptom scores, the SHT group and the controls, and the males and the females, were compared with the Mann-Whitney U test. The Kruskal-Wallis test was used to evaluate other predictors of emotional function and symptom scores. For this purpose, the subjects were divided into age, BMI, TSH, free T4, and free T3 quartiles. The Mann-Whitney U test was used as post hoc test between the lowest and the highest quartiles. The main analyses were done comparing the SHT group with TSH range 3.5–10.0 mIU/liter with the controls, but in addition, those in the SHT group with serum TSH in the range 5.0 –10.0 mIU/liter were also compared separately with the control group. In the intervention study, comparisons between the T4 and placebo groups at the start and end of the intervention were done similarly as between the SHT group and the controls. In addition, ⌬ values (value at end of intervention minus value at inclusion) were compared between the two groups with similar statistics. Correlations were evaluated with the Pearson correlation coefficient. Unless otherwise stated, all data are expressed as mean ⫾ sd. All tests were done two-sided, and P ⬍ 0.05 was considered statistically significant. Corrections for multiple comparisons were not performed. Statistical analyses were performed with SPSS version 11.0 (SPSS Inc., Chicago, IL).

147

Ethics The Regional Ethics Committee approved the study, and all participants gave their written informed consent.

Results

Among the 7954 subjects with serum TSH measurements in the fifth Tromsø study, 1253 were excluded because of the health questionnaire (heart disease, stroke, or diabetes) and an additional 576 subjects had answered no to the question on willingness to participate in follow-up studies. Among the remaining 6125 subjects, 363 had serum TSH in the range 3.5–10.0 mIU/liter. After review of hospital journals, 114 were excluded because of T4 use, illness, or participation in other studies. The remaining 249 subjects were invited to the follow-up study; 167 attended, and 89 fulfilled the criteria for SHT. Among these, 38 subjects had serum TSH in the 5.0 – 10.0 mIU/liter range. Also, 249 age- and sex-matched controls who in the fifth Tromsø study had serum TSH in the range 0.50 –3.49 mIU/liter were invited; 162 attended, and 154 had normal serum free T4 and free T3 and serum TSH in the same range as in the fifth Tromsø study at the reexamination. The characteristics of the SHT and control subjects are given in Table 1. Tests of cognitive function

There were no significant differences between the SHT group and the controls on any of the individual tests or on the composite score for cognitive function (Table 2). In the multiple linear regression model, age was the most important (and negative) predictor of cognitive performance. The serum TSH level was significantly and negatively associated with performance on the Trail Making test A; the serum free T4 level was significantly and positively associated with performance on the Stroop test parts 1 and 2 and the word association test; whereas the serum free T3 level was significantly and negatively associated with performance at the visual recall test (Table 3). There was no significant association between serum TSH, free T4, or free T3 and the composite score for cognitive function (Table 3). Tests of emotional function

Those with SHT had a significantly more favorable score than the controls on the GHQ-30 test (P ⫽ 0.006, MannWhitney U test) (Table 4). There was an almost significant

TABLE 1. Characteristics of the subjects in the SHT and control groups and those in the intervention study All subjects

Males/females Age (yr) BMI (kg/m2) Smokers/nonsmokers Serum TSH (mIU/liter) Serum free T4 (pmol/liter) Serum free T3 (pmol/liter) a

P ⬍ 0.05;

b

Nonsmokers

SHT group with serum TSH 3.5–10 mIU/liter

SHT group with serum TSH 5.0 –10 mIU/liter

Control group



45/44 62.4 ⫾ 11.9 28.0 ⫾ 5.1a 11/78b 5.72 ⫾ 1.68b 12.7 ⫾ 1.9b 3.7 ⫾ 0.7

26/12a 62.5 ⫾ 11.7 27.8 ⫾ 4.4 8/38 6.45 ⫾ 1.34b 12.7 ⫾ 1.9b 3.8 ⫾ 0.8

72/82 61.0 ⫾ 12.5 26.4 ⫾ 3.8 45/109 1.79 ⫾ 0.69 14.0 ⫾ 2.1 3.9 ⫾ 0.8

39/39 62.0 ⫾ 12.3 28.3 ⫾ 5.2

21/9a 61.7 ⫾ 12.3 28.2 ⫾ 4.4

5.51 ⫾ 1.33b 6.25 ⫾ 1.28b 12.8 ⫾ 1.7b 12.8 ⫾ 1.4b 3.7 ⫾ 0.7 3.9 ⫾ 0.8

P ⬍ 0.01, SHT vs. control group (Student’s t test or ␹2 test).

Intervention study Control group

T4 group

48/61 19/17 61.3 ⫾ 12.6 61.6 ⫾ 11.5 27.1 ⫾ 3.8 28.6 ⫾ 5.4 6/30 1.83 ⫾ 0.71 5.81 ⫾ 1.76 13.9 ⫾ 2.2 12.7 ⫾ 1.8 3.8 ⫾ 0.7 3.6 ⫾ 0.6

Placebo group

18/15 63.0 ⫾ 12.4 27.0 ⫾ 3.7 3/30 5.32 ⫾ 1.25 13.0 ⫾ 2.1 3.8 ⫾ 0.7

b

30 28 29 29 29 30 30

28 28 28 30 28

48.7 ⫾ 20.0 50.8 ⫾ 18.8 42.8 ⫾ 13.5 20.2 ⫾ 5.6 16.8 ⫾ 6.4 24.4 ⫾ 6.1 37.1 ⫾ 13.3

68.1 ⫾ 39.0 105 ⫾ 51 1603 ⫾ 304 18.2 ⫾ 5.0 ⫺0.3 ⫾ 4.6

78 76 77 77 77 78 78

75 75 75 78 75

2 2 1

1 1 1

1

2 2

2

1 1

Arrow indicates favorable (1) or unfavorable (2) result. Immediate and delayed recall.

30 30 28

5.6 ⫾ 1.0 4.1 ⫾ 1.0 388 ⫾ 114

n

78 78 75

Score

1 1 2

n

18.9 ⫾ 4.9 0.2 ⫾ 4.4

1632 ⫾ 286

67.1 ⫾ 28.4 99 ⫾ 46

38.6 ⫾ 12.4

21.4 ⫾ 5.5 17.9 ⫾ 6.1 24.3 ⫾ 5.6

51.5 ⫾ 21.2 51.2 ⫾ 11.5 43.4 ⫾ 12.0

5.9 ⫾ 1.3 4.2 ⫾ 1.0 384 ⫾ 131

Score


107 102

102

103 103

104

104 104 104

105 103 104

107 107 102

n

18.3 ⫾ 4.9 0.7 ⫾ 4.2

1573 ⫾ 324

62.6 ⫾ 21.7 101 ⫾ 57

39.1 ⫾ 14.0

21.9 ⫾ 5.3 16.4 ⫾ 7.4 26.1 ⫾ 7.0

44.0 ⫾ 15.7 48.0 ⫾ 9.1 43.7 ⫾ 13.3

5.6 ⫾ 1.1 4.3 ⫾ 1.2 395 ⫾ 159

Score

Control group

36 34

34

34 36

36

34 36 36

36 35 36

36 36 34

n

18.9 ⫾ 3.7 1.8 ⫾ 3.4

1573 ⫾ 313

63.1 ⫾ 21.0 92 ⫾ 44

37.9 ⫾ 12.4

21.4 ⫾ 6.1 177 ⫾ 6.6 25.8 ⫾ 5.7

45.5 ⫾ 17.0 48.0 ⫾ 9.0 45.7 ⫾ 12.1

5.8 ⫾ 1.0 4.1 ⫾ 0.9 424 ⫾ 236

Score

T4 group at baseline

33 31

32

31 31

33

33 33 33

33 31 33

33 33 32

n

18.1 ⫾ 5.4 ⫺1.1 ⫾ 4.7

1633 ⫾ 348

74.3 ⫾ 54.4 111 ⫾ 48

35.0 ⫾ 12.1

19.8 ⫾ 5.2 16.6 ⫾ 6.3 23.7 ⫾ 6.8

50.0 ⫾ 19.7 56.2 ⫾ 27.7 39.5 ⫾ 14.2

5.5 ⫾ 1.2 4.2 ⫾ 1.4 399 ⫾ 132

Score

Placebo group at baseline

36 35

35

36 36

36

35 36 35

36 36 36

36 36 35

n

19.8 ⫾ 3.5 1.5 ⫾ 3.7

1483 ⫾ 302

61.1 ⫾ 27.3 94 ⫾ 62

41.7 ⫾ 15.0

23.3 ⫾ 5.8 19.7 ⫾ 7.5 26.3 ⫾ 5.6

39.0 ⫾ 14.8 47.1 ⫾ 8.9 48.2 ⫾ 14.4

5.8 ⫾ 1.0 4.2 ⫾ 0.9 372 ⫾ 108

Score

T4 group at 12 months

Intervention study

33 30

33

31 30

33

33 33 33

33 31 33

33 33 33

n

19.7 ⫾ 4.2 ⫺0.9 ⫾ 4.8

1562 ⫾ 331

66.7 ⫾ 28.2 103 ⫾ 49

41.6 ⫾ 14.1

22.7 ⫾ 5.5 16.9 ⫾ 7.9 25.2 ⫾ 9.5

44.1 ⫾ 17.7 50.7 ⫾ 13.2 40.7 ⫾ 15.8

5.8 ⫾ 1.3 4.1 ⫾ 1.1 409 ⫾ 114

Score

Placebo group at 12 months


a

Attention and working memory Digit Span forward Digit Span backward Seashore Rhythm test Psychomotor/cognitive speed Trail Making test A Stroop test, parts 1 and 2 Digit Symbol test Memory Verbal recallb Visual recallb Word list test Language Controlled word association test Cognitive flexibility/executive function Stroop test, part 3 Trail Making test B Speed of information processing CalCAP Intelligence Vocabulary (WAIS) Composite cognitive score

Resulta


TABLE 2. Score on cognitive function tests in SHT and control groups (nonsmokers) and at baseline and after 12 months in the intervention study

148 Jorde et al. • Neuropsychological Function and TSH



149

TABLE 3. Standardized ␤-coefficients and t values from the linear regression models in the 177 nonsmoking subjects who completed all cognitive function tests Independent variables Dependent variables

Resulta

Genderb

␤

Attention and working memory Digit Span forward Digit Span backward Seashore Rhythm test Psychomotor/cognitive speed Trail Making test A Stroop test, parts 1 and 2 Digit Symbol test Memory Verbal recalld Visual recalld Word list test Language Word association test Cognitive flexibility/executive function Stroop test, part 3 Trail Making test B Speed of information processing CalCAP Intelligence Vocabulary (WAIS) Composite cognitive score

t

Age

␤

BMI

␤

t

TSH t

␤

t

␤

R2

Free T3c

Free T4 t

␤

t

1 1 2

⫺0.05 ⫺0.54 ⫺0.11 ⫺0.00 ⫺0.02 ⫺0.21 ⫺0.10 ⫺1.14 0.18

⫺1.52 ⫺0.05 ⫺0.64 0.12 1.50 0.10 1.15 ⫺0.14 ⫺1.62 0.04 ⫺2.89 ⫺0.11 ⫺1.56 ⫺0.05 ⫺0.68 0.00 0.05 ⫺0.03 ⫺0.33 0.07 2.40 0.01 0.09 ⫺0.07 ⫺0.90 ⫺0.12 ⫺1.42 0.07 0.73 0.05

2 2 1

0.09 1.19 0.35 5.09 0.04 0.51 0.19 2.62 0.05 0.70 0.12 1.43 0.17 ⫺0.29 ⫺3.88 0.37 5.48 0.10 1.52 ⫺0.01 ⫺0.07 ⫺0.16 ⫺2.20 ⫺0.00 ⫺0.04 0.26 0.24 3.65 ⫺0.58 ⫺10.1 ⫺0.14 ⫺2.42 0.07 1.11 0.12 1.80 ⫺0.09 ⫺1.26 0.44

1 1 1

0.12 0.09 0.13

1.55 ⫺0.34 1.23 ⫺0.57 1.88 ⫺0.48

⫺4.99 ⫺0.06 ⫺0.84 ⫺0.12 ⫺1.58 ⫺0.11 ⫺1.45 ⫺0.08 ⫺0.94 0.20 ⫺9.36 ⫺0.11 ⫺1.71 0.07 1.05 0.07 0.97 ⫺0.15 ⫺2.15 0.37 ⫺7.50 ⫺0.13 ⫺1.97 ⫺0.05 ⫺0.80 0.04 0.63 ⫺0.11 ⫺1.54 0.31

1

0.23

2.81 ⫺0.07

⫺0.89 ⫺0.08 ⫺1.06

2 2

⫺0.14 ⫺2.10 ⫺0.05 ⫺0.64

0.05

0.65

0.23

2.85

0.02

0.17 0.08

0.58 0.50

9.71 7.49

0.12 0.02

2.04 ⫺0.03 ⫺0.44 ⫺0.13 ⫺1.91 0.33 0.05 0.66 ⫺0.01 ⫺0.08

0.02 0.06

0.28 0.41 0.76 0.27

0.32

4.40

0.11

1.53

0.27 ⫺0.03 ⫺0.44

0.08

0.95 0.13

2

0.06

0.77

1 1

0.08 0.12

0.92 ⫺0.16 1.71 ⫺0.51

0.02

⫺2.22 ⫺0.12 ⫺1.59 0.04 0.06 ⫺7.89 ⫺0.14 ⫺2.17 ⫺0.00 ⫺0.03

0.05 0.11

0.60 ⫺0.03 ⫺0.30 0.05 1.54 ⫺0.10 ⫺1.37 0.34

R2 values are based on the model with gender, age, BMI, serum TSH, and free T4 as independent variables. Values of 兩t兩 ⬎ 1.96, 兩t兩 ⬎ 2.58, and 兩t兩 ⬎ 3.29 correspond to P ⬍ 0.05, P ⬍ 0.01, and P ⬍ 0.001, respectively. a Arrow indicates favorable (1) or unfavorable (2) result. b Male ⫽ 1; female ⫽ 2. c When free T3 was used in the model, the other covariables were gender, age, BMI, and serum TSH. d Immediate and delayed recall.

difference between males and females on the GHQ-30 test (1.6 ⫾ 3.1 and 2.5 ⫾ 4.2, respectively, P ⫽ 0.07, MannWhitney U test). Therefore, the GHQ-30 scores in the SHT and control group were compared in males and females separately [1.0 ⫾ 2.1 vs. 1.9 ⫾ 3.5 (P ⫽ 0.03) in males and 1.2 ⫾ 2.0 vs. 3.2 ⫾ 4.9 in females (P ⫽ 0.06), respectively, Mann-Whitney U test]. When the cohort was divided into serum TSH quartiles, there was a significant association

between serum TSH and the GHQ-30 score (P ⫽ 0.03, Kruskal-Wallis test). The GHQ-30 score in the lowest and highest serum TSH quartiles were 3.4 ⫾ 5.4 and 1.4 ⫾ 2.4, respectively (P ⫽ 0.05, Mann-Whitney U test). Also, when evaluating nonsmokers separately, there was a significant difference between the SHT and control group regarding the GHQ-30 test (Table 4). On the BDI, those with SHT had more favorable results

TABLE 4. Score on emotional function tests in SHT and control groups All subjects SHT group with serum TSH 3.5–10 mIU/liter

Resulta

Depressed mood BDI (1–13) BDI (14 –21) BDI total score Mental health status GHQ-30 Likert score GHQ factor scores A (anxiety) B (feeling of incompetence) C (depression, hopelessness) D (difficulty in coping) E (social dysfunction) a b

Nonsmokers


Control group

n

2 2 2

88 88 88

1.7 ⫾ 2.0 2.2 ⫾ 2.1 3.9 ⫾ 3.7

38 38 38

2 2

88 1.1 ⫾ 2.1b 38 1.4 ⫾ 2.6 145 2.6 ⫾ 4.3 78 1.1 ⫾ 2.1b 30 1.6 ⫾ 2.9 104 2.7 ⫾ 4.5 88 22.2 ⫾ 5.5 38 22.9 ⫾ 5.9 145 24.2 ⫾ 8.7 78 22.3 ⫾ 5.5 30 23.3 ⫾ 6.2 104 24.2 ⫾ 9.1

2 2 2 2 2

88 88 88 88 88

Arrow indicates favorable (1) or unfavorable (2) result. P ⬍ 0.01, SHT vs. control group (Mann-Whitney U test).

1.9 ⫾ 2.4 148 2.2 ⫾ 2.1 148 4.0 ⫾ 4.2 148

4.8 ⫾ 2.9 4.1 ⫾ 0.7 2.2 ⫾ 1.5 3.9 ⫾ 1.2 3.0 ⫾ 0.9

145 145 145 145 145

Score

n

2.5 ⫾ 2.9 78 2.7 ⫾ 2.3 78 5.1 ⫾ 4.7 78

5.5 ⫾ 4.1 4.2 ⫾ 0.9 2.0 ⫾ 1.6 4.2 ⫾ 1.5 3.1 ⫾ 1.2

78 78 78 78 78

Score

n

1.7 ⫾ 2.0 2.1 ⫾ 2.1 3.8 ⫾ 3.7

30 30 30

4.7 ⫾ 2.8 4.0 ⫾ 0.7 2.0 ⫾ 1.5 3.8 ⫾ 1.1 2.9 ⫾ 1.0

30 30 30 30 30

Score

Control group

Score

38 38 38 38 38

n


n

4.7 ⫾ 2.8 3.9 ⫾ 0.7 2.0 ⫾ 1.4 3.8 ⫾ 1.1 2.9 ⫾ 1.0

Score


n

1.8 ⫾ 2.5 107 2.1 ⫾ 2.1 107 3.9 ⫾ 4.3 107

4.9 ⫾ 3.1 4.2 ⫾ 0.7 2.2 ⫾ 1.5 4.1 ⫾ 1.3 2.9 ⫾ 0.9

104 104 104 104 104

Score

2.4 ⫾ 2.8 2.6 ⫾ 2.2 5.0 ⫾ 4.6

5.6 ⫾ 4.2 4.1 ⫾ 0.9 2.0 ⫾ 1.6 4.2 ⫾ 1.6 3.1 ⫾ 1.3

150



than the controls, but the differences were not significant. Contrary to that seen for cognitive function, there was no association between age and emotional function. Questionnaire on hypothyroid symptoms

Except for a single question (“Are you more tired than before?”) where those in the SHT group with serum TSH in the range 5.0 –10.0 mIU/liter scored significantly more favorably than the controls, the SHT group and the controls did not differ significantly on any of the questions on hypothyroid symptoms or on the total symptom score (Table 5). Examining the nonsmokers separately did not add additional information. There was no significant relation between quartiles of serum TSH, free T4, or free T3 and the total symptom score when evaluated with the Kruskal-Wallis test. Intervention study

Thirty-six subjects were included in the T4 group and 34 subjects in the placebo group. One subject in the placebo group dropped out after 6 months because of serious disease unrelated to thyroid function and was excluded from analyses in the intervention study. Characteristics of the subjects are given in Table 1. In the T4 group, one subject had serum TSH less than 0.05 mIU/liter, nine subjects had 0.05– 0.48 mIU/liter, nine subjects had 0.53–1.37 mIU/liter, nine subjects had 1.52–1.93 mIU/liter, four subjects had 2.17–2.74 mIU/liter, and four subjects had 3.75– 6.85 mIU/liter at the end of the study. The compliance rate in the T4 group and the placebo group for the entire study was 0.91 ⫾ 0.06 in both groups. There was a significant and negative correlation between compliance rate the first 3 months and serum TSH at the 3-month visit (r ⫽ ⫺0.40; P ⫽ 0.017) but not at later

intervals. The thyroid status, T4 use, and compliance during the intervention study are given in Table 6. There were no significant differences in any parameter on cognitive function (Table 2), emotional function (Table 7), or hypothyroid symptoms (Table 8) between the T4 and the placebo group, either at the start of the intervention or at the end of the study. The symptom score at the end of the study (10 questions on change of symptom) were ⫺0.08 ⫾ 2.23 and ⫺0.65 ⫾ 2.04 in the T4 and placebo groups, respectively. Furthermore, when evaluating ⌬ values (scores at the end of study minus scores at the start), there were no significant differences regarding cognitive or emotional function, whereas on one of the individual questions regarding change in hypothyroid symptoms (“Are your clothes looser/no change/tighter than one year before?”), those given T4 scored significantly more in the positive direction (clothes looser) than the placebo group (Table 8). There was no relation between these ⌬ values and ⌬ serum TSH, ⌬ free T4, and ⌬ free T3, either when looking at all 69 subjects in the intervention study together or when analyzing those given T4 and those given placebo separately. At the end of the study, most of the subjects felt no general improvement in health, and most subjects in both groups thought they had been given placebo or answered “do not know” (Table 8). Discussion

In the present study, we have found no significant differences in cognitive function or hypothyroid symptoms between the SHT and the control group, whereas for emotional function, those with SHT scored more favorably than the controls on the GHQ-30. In the SHT group, we included subjects with serum TSH

TABLE 5. Percentage of subjects who answered yes to questions related to hypothyroidism in the SHT and control groups and the total symptom score in the SHT and control groups All subjects

Have you gained weight in the last year? Do you feel colder than other people? Is your skin more dry now than 1 yr ago? Are your clothes tighter than 1 yr ago? Is your hair drier than 1 yr ago? Are your hands puffier than before? Are your muscles weaker than before? Are your nails more brittle than before? Do you feel “pins and needles” in your hands or feet? Do you fatigue easier than 1 yr ago? Do you need to dress more warmly than others? Do you have muscle cramps? Are you more short of breath than 1 yr ago? Do your feet swell? Is your voice coarser than before? Is your skin colder than before? Do you have numbness in hands or feet? Are you more constipated than 1 yr ago? Are you more tired than before? Total symptom score [median (range)] a

Nonsmokers

SHT group with TSH 3.5–10 mIU/liter (n ⫽ 89)

SHT group with TSH 5.0 –10 mIU/liter (n ⫽ 38)

Control group (n ⫽ 154)

SHT group with TSH 3.5–10 mIU/liter (n ⫽ 78)

SHT group with TSH 5.0 –10 mIU/liter (n ⫽ 30)

Control group (n ⫽ 109)

27.0 32.6 37.1 25.8 23.6 18.0 28.1 25.8 19.1 39.3 25.8 25.8 23.6 21.3 22.5 15.7 18.0 5.6 34.5 4 (0 –14)

23.7 36.8 28.9 23.7 21.1 10.5 26.3 21.1 10.5 42.1 28.9 23.7 23.7 15.8 21.1 15.8 15.8 7.9 18.4a 3 (0 –13)

34.4 35.7 30.5 35.1 20.8 20.8 31.8 22.1 19.5 42.9 21.4 23.5 26.0 21.4 22.1 18.8 15.6 9.1 41.8 4 (0 –18)

26.9 33.3 38.5 28.2 25.6 17.9 29.5 21.8 19.2 37.2 25.6 25.6 21.8 24.3 21.8 14.1 19.2 5.1 35.9 4 (0 –14)

23.3 36.7 30.0 26.7 20.0 10.0 23.3 13.3 10.0 33.3 26.7 26.7 20.0 20.0 16.7 10.0 13.3 6.7 16.7a 3 (0 –13)

37.6 33.9 27.5 37.6 19.3 20.2 27.5 25.7 19.3 36.7 18.3 20.2 24.8 18.3 22.9 14.7 13.8 10.1 42.2 4 (0 –14)

P ⬍ 0.01, SHT vs. control group (Mann-Whitney U test).



151

TABLE 6. T4 dose and compliance rate in those given T4, and serum TSH, free T4, and free T3 in the T4 and placebo groups at inclusion and at 3-month intervals during the intervention study Placebo group (n ⫽ 33)

T4 group (n ⫽ 36) Compliance rate

TSH (mIU/liter)

Free T4 (pmol/liter)


TSH (mIU/liter)



0.95 ⫾ 0.01 0.90 ⫾ 0.09 0.91 ⫾ 0.09 0.91 ⫾ 0.10

5.81 ⫾ 1.76 0.95 ⫾ 0.98 1.30 ⫾ 1.66 2.26 ⫾ 1.78 1.52 ⫾ 1.51

12.7 ⫾ 1.8 19.6 ⫾ 3.2 17.7 ⫾ 2.6 16.7 ⫾ 2.5 18.0 ⫾ 2.8

3.60 ⫾ 0.63 4.54 ⫾ 0.86 4.09 ⫾ 0.67 4.21 ⫾ 0.63 4.43 ⫾ 0.58

5.32 ⫾ 1.25 5.14 ⫾ 1.61 5.14 ⫾ 2.56 5.58 ⫾ 1.64 5.42 ⫾ 1.96

13.0 ⫾ 2.1 13.5 ⫾ 2.2 14.1 ⫾ 2.2 13.4 ⫾ 2.3 13.9 ⫾ 2.3

3.78 ⫾ 0.70 4.35 ⫾ 0.70 4.13 ⫾ 0.69 4.53 ⫾ 0.68 4.65 ⫾ 0.70

T4 dose (␮g/d)

Inclusion 3 months 6 months 9 months 12 months

50/100 97.2 ⫾ 23.0 94.4 ⫾ 28.1 109.7 ⫾ 24.8

3.5–10.0 mIU/liter, which is a lower range than what has been used in most other studies. However, the upper normal TSH level is a matter of definition and depends on whether subjects with presence of thyroid autoantibodies and other risk factors for hypothyroidism are included or not. Thus, in the study by Bjøro et al. (29) the upper normal serum TSH level was approximately 3.5 mIU/liter if excluding those with thyroid autoantibodies, and in the NHANES III study (30) the TSH 97.5 percentile was reduced from 5.80 mIU/liter in the total population to 4.12 mIU/liter if excluding those with thyroid autoantibodies and other risk factors. Cognitive symptoms such as slow mentation, reduced memory function, and inability to concentrate are frequently reported in subjects with overt hypothyroidism (31, 32) and also in some studies on SHT. Thus, impaired memory function in SHT was found by Baldini et al. (33), del Ser Quijano et al. (34), and Monzani et al. (10). In the latter study, 14 subjects with SHT were compared with 50 control subjects. There was a significant impairment in memory in the SHT group and a significant improvement after treatment with T4. However, their group of SHT subjects had a mean serum TSH of 8.8 mIU/liter, which was considerably higher than in our study. Similarly, Jaeschke et al. (12) found T4 treatment to significantly improve a composite psychometric memory score in SHT, but their mean serum TSH level was 12.1 mIU/liter, which in this context is rather high. In our study, we applied a broad range of tests of cognitive function. On 11 of the 14 tests, the control subjects scored slightly better than the SHT group, and they also had a better composite cognitive function score. However, none of the differences were statistically significant. Furthermore, in the

multiple linear regression model, the direction of the association between serum TSH and cognitive function was in favor of a low serum TSH in only six of the 14 tests, and on only one of these six tests (the Trail Making A test) was the association statistically significant. Because we did not do a Bonferroni or other correction for multiple testing, this significance could well be the result of chance. Accordingly, cognitive function was not markedly affected in our SHT group, probably because their serum TSH levels were not raised above 10 mIU/liter. Similar to cognitive function, there is an association between overt hypothyroidism and depression (35). Depressive features are also reported more frequently in SHT with slight improvement after T4 treatment (10, 36). However, in a large epidemiological study by Pop et al. (37), depression was only weakly and nonsignificantly associated with SHT, and in our study the subjects with SHT actually had a better score on the BDI than the controls. Furthermore, in our study those with SHT had a significantly better score than the controls on the GHQ-30. When evaluated with the Likertscoring method, this trend was also seen for each of the five subfactors anxiety, feeling of incompetence, depression, difficulty in coping, and social dysfunction. The classical signs of hypothyroidism are well known, and symptom scores have been developed that clearly discriminates between euthyroid and hypothyroid subjects (38, 39). The number of symptoms correlate with degree of hypothyroidism (38, 39), which was also demonstrated in the large epidemiological Colorado Thyroid Disease Prevalence Study (40) that included 25,862 subjects voluntarily attending a health fair. Among the 2336 subjects classified as SHT (in-

TABLE 7. Score on emotional function tests at baseline and after 12 months in those given T4 and placebo in the intervention study Resulta

T4 group at baseline n

Depressed mood BDI (1–13) BDI (14 –21) BDI total score Mental health status GHQ-30 Likert score GHQ factor scores A (anxiety) B (feeling of incompetence) C (depression, hopelessness) D (difficulty in coping) E (social dysfunction) a

Score

Placebo group at baseline n

Score

T4 group at 12 months n

Score

Placebo group at 12 months n

Score

2 2 2

36 36 36

1.9 ⫾ 2.0 2.4 ⫾ 2.2 4.4 ⫾ 3.7

33 33 33

1.6 ⫾ 2.1 2.0 ⫾ 1.9 3.7 ⫾ 3.8

36 36 36

2.0 ⫾ 2.0 2.3 ⫾ 2.1 4.3 ⫾ 3.6

33 33 33

1.9 ⫾ 2.4 1.4 ⫾ 2.0 3.3 ⫾ 4.0

2 2

36 36

1.5 ⫾ 2.3 22.6 ⫾ 5.5

33 33

0.7 ⫾ 1.3 21.6 ⫾ 5.2

36 36

1.9 ⫾ 3.3 23.9 ⫾ 7.2

32 32

1.2 ⫾ 2.0 23.5 ⫾ 6.1

2 2 2 2 2

36 36 36 36 36

4.8 ⫾ 2.9 4.0 ⫾ 0.7 1.8 ⫾ 1.2 3.8 ⫾ 1.2 3.1 ⫾ 1.0

33 33 33 33 33

4.6 ⫾ 2.5 3.9 ⫾ 0.7 2.1 ⫾ 1.6 3.6 ⫾ 0.8 2.8 ⫾ 1.1

36 36 36 36 36

5.7 ⫾ 3.1 4.0 ⫾ 0.9 2.1 ⫾ 1.6 4.0 ⫾ 1.4 3.1 ⫾ 1.2

32 32 32 32 32

5.5 ⫾ 3.0 4.0 ⫾ 0.6 2.2 ⫾ 1.6 4.0 ⫾ 1.2 3.0 ⫾ 0.5

Arrow indicates favorable (1) or unfavorable (2) result.

152



TABLE 8. Change in symptoms at end of the intervention study and belief concerning drug given in the T4 and placebo groups

If comparing with 1 yr ago, do you feel warmer/no change/colder? is your skin less dry/no change/more dry? are your clothes looser/no change/tighter? is your hair less dry/no change/more dry? are your muscles stronger/no change/weaker? are your nails less brittle/no change/more brittle? are you less short of breath/no change/more short of breath? is your voice less coarse/no change/more coarse? is your skin warmer/no change/colder? do you feel less constipated/no change/more constipated? do you feel better/no change/worse? Do you believe you were given T4/placebo/do not know? a

T4 group

Placebo group

6/27/3 4/25/7 10/24/2a 3/25/8 0/32/4 3/29/4 2/31/3 1/30/5 3/32/1 6/26/4 4/29/3 3/15/18

4/26/3 1/27/5 4/22/7 1/28/4 2/26/5 1/26/5 2/25/6 1/25/6 3/28/2 2/30/1 3/25/5 4/10/18

P ⬍ 0.05, T4 vs. placebo group (Mann-Whitney U test).

cluding 269 subjects taking T4), the percentage of subjects reporting symptoms associated with hypothyroidism was slightly but significantly higher than in the euthyroid group (13.7 vs. 12.1%) (40). Furthermore, in a study of 63 patients with SHT treated with T4 or placebo by Meier et al. (11), those given T4 had a significant improvement in symptom score as compared with baseline values. However, the change in symptoms over the 48-wk treatment period between the placebo group and the T4 group did not differ significantly, and the improvement in symptom score was seen in those that had serum TSH higher than 12 mIU/liter before the start of T4 treatment. In accordance with this, because we included only subjects in the SHT group that had serum TSH between 3.5 and 10.0 mIU/liter, there was no difference between the SHT and control group regarding symptoms in our study. In addition to the comparison between the SHT and control group, we also performed a double-blind placebo-controlled intervention study with T4 for 1 yr. Because we did not find any significant differences between the SHT group and the controls (except for GHQ-30), it was no surprise that intervention with T4 had no effect on cognitive function or depression and that most of the subjects thought they had received placebo. Furthermore, there was no negative effect on the GHQ-30 score by T4 substitution, which does bring into question the importance of our finding of a difference between the SHT and control group regarding this score. We tried to tailor the T4 dose in steps of 25 ␮g to achieve a serum TSH between 0.5 and 1.5 mIU/liter. However, despite acceptable compliance and strict adherence to a dosage algorithm designed to reach this target, and serum TSH measurement and dose adjustment every third month, the treatment goal was reached only in nine of the 36 subjects. This illustrates that a target range of 0.50 –1.50 mIU/liter probably is too narrow and that T4 doses may need to be titrated in smaller steps than 25 ␮g per day. However, there was no relation between change in serum TSH in the intervention group and any of the parameters measured, and because the average serum TSH was 1.52 mIU/liter at the end of the intervention, we feel it is unlikely that a slightly higher average T4 dose would have significantly affected the neuropsychological test scores. Our study has several weaknesses. We cannot rule out that we would have found significant differences between the SHT and the control group had we included more subjects.

However, to include 89 subjects with SHT, we had to screen almost 8000 subjects, and a larger group would be hard to find. Furthermore, including subjects from an epidemiological survey would exclude those whose hypothyroid symptoms had made them seek medical help leading to T4 treatment. Accordingly, the true prevalence of symptoms and dysfunction related to serum TSH in the range 3.5–10.0 mIU/ liter and with free T4 and free T3 within the reference range, is probably higher than that found by us. Furthermore, our SHT patients had a mean TSH level of 5.5 mIU/liter, and our results do not necessarily apply to those with higher serum TSH levels. However, when looking separately at those with serum TSH in the range 5.0 –10.0 mIU/liter, there was no trend toward a difference between these SHT subjects and the controls. The subjects were not informed about their thyroid status when they were invited, but they knew that the focus of the study was thyroid dysfunction. Because only 66% of those invited to the follow-up examination attended, we cannot rule out a selection basis favoring those with symptoms compatible with thyroid disease. In addition, we excluded a considerable number of subjects because of concomitant diseases, and our results do therefore mostly apply to a fairly healthy population. The SHT and control groups were well matched regarding age and gender. However, there were significantly more smokers in the control group than in the SHT group. This could possibly be explained by the TSH-lowering effect of smoking (41), which would reduce the number of smokers with TSH values slightly above our 3.5 mIU/liter limit. Because of this difference in number of smokers, and the interactions between smoking and TSH status regarding cognitive function, subanalyses in nonsmokers were done on all tests. However, the number of smokers was too small to analyze these subjects separately, and we cannot exclude an effect of SHT on neuropsychological function in these subjects. In the intervention study, the mean serum TSH after 1 yr in those given T4 was 1.52 mIU/liter, which was approximately 0.5 mIU/liter above the target level. However, this is lower than in most other studies on the effects of T4 on neuropsychological function and symptoms (11–13), and we find it unlikely that a higher dose of T4 would have disclosed differences between those treated and those not. On the other hand, our study has considerable strength. We applied a broad range of tests and used strict selection


criteria for both SHT and control subjects. Those included were recruited from an epidemiological survey and not from clinical practice, which would have favored inclusion of subjects with symptoms unrelated to thyroid diseases. Our subjects also had a stable thyroid function because the serum TSH level was elevated (or normal) on at least two occasions before inclusion in the study. These points on inclusion criteria for SHT subjects in clinical trials have recently been emphasized in a commentary by Biondi et al. (42). Furthermore, we had a serum TSH level of 10.0 mIU/liter as the upper limit for inclusion in the SHT group. Patients with higher TSH levels may well have subclinical disease, but most of these patients will be started on treatment regardless of symptoms. Finally, although it could be considered a weakness that our patients had SHT with serum TSH in the narrow range 3.5–10 mIU/liter, this is also a strength of the study and shows that, at least with regard to cognitive and emotional function, there is no need to start treatment because of a slightly elevated serum TSH level. In conclusion, in subjects whose serum TSH levels are not raised above 10 mIU/liter, as in our study, there are hardly any symptoms or neuropsychological dysfunctions. Acknowledgments The assistance by the staff at the Clinical Research Unit, University Hospital of North Norway, and by test technician Kari Bjerkaas, Neuropsychological Laboratory, Department of Neurology, University Hospital of North Norway, is gratefully acknowledged. The T4 and placebo tablets were generously supplied by NycoMed Pharma. Received August 8, 2005. Accepted October 25, 2005. Address all correspondence and requests for reprints to: Rolf Jorde, Medical Department B, University Hospital of North Norway, 9038 Tromsø, Norway. E-mail: [email protected]. This work was supported by a grant from the Norwegian Research Council and The Northern Norway Regional Health Authority.

References 1. McDermott MT, Ridgway EC 2001 Subclinical hypothyroidism is mild thyroid failure and should be treated. J Clin Endocrinol Metab 86:4585– 4590 2. Chu JW, Crapo LM 2001 The treatment of subclinical hypothyroidism is seldom necessary. J Clin Endocrinol Metab 86:4591– 4599 3. Owen PJ, Lazarus JH 2003 Subclinical hypothyroidism: the case for treatment. Trends Endocrinol Metab 14:257–261 4. Vanderpump M 2003 Subclinical hypothyroidism: the case against treatment. Trends Endocrinol Metab 14:262–266 5. Ayala AR, Danese MD, Ladenson PW 2000 When to treat mild hypothyroidism. Endocrinol Metab Clin North Am 29:399 – 415 6. Fatourechi V 2001 Subclinical thyroid disease. Mayo Clin Proc 76:413– 416 7. Cooper DS 2001 Subclinical hypothyroidism. N Engl J Med 345:260 –265 8. Surks MI, Ortiz E, Daniels GH, Sawin CT, Col NF, Cobin RH, Franklyn JA, Hershman JM, Burman KD, Denke MA, Gorman C, Cooper RS, Weissman NJ 2004 Subclinical thyroid disease: scientific review and guidelines for diagnosis and management. JAMA 291:228 –238 9. Danese MD, Ladenson PW, Meinert CL, Powe NR 2000 Effect of thyroxine therapy on serum lipoproteins in patients with mild thyroid failure: a quantitative review of the literature. J Clin Endocrinol Metab 85:2993–3001 10. Monzani F, Del Guerra P, Caraccio N, Pruneti CA, Pucci E, Luisi M, Baschieri L 1993 Subclinical hypothyroidism: neurobehavioral features and beneficial effect of l-thyroxine treatment. Clin Investig 71:367–371 11. Meier C, Staub JJ, Roth CB, Guglielmetti M, Kunz M, Miserez AR, Drewe J, Huber P, Herzog R, Mu¨ller B 2001 TSH-controlled l-thyroxine therapy reduces cholesterol levels and clinical symptoms in subclinical hypothyroidism: a double blind, placebo-controlled trial (Basel thyroid study). J Clin Endocrinol Metab 86:4860 – 4866


153

12. Jaeschke R, Guyatt G, Gerstein H, Patterson C, Molloy W, Cook D, Harper S, Griffith L, Carbotte R 1996 Does treatment with l-thyroxine influence health status in middle-aged and older adults with subclinical hypothyroidism? J Gen Intern Med 11:744 –749 13. Kong WM, Sheikh MH, Lumb PJ, Freedman DB, Crook M, Dore´ CJ, Finer N 2002 A 6-month randomized trial of thyroxine treatment in women with mild subclinical hypothyroidism. Am J Med 112:348 –354 14. Pollock MA, Sturrock A, Marshall K, Davidson KM, Kelly CJG, McMahon AD, McLaren EH 2001 Thyroxine treatment in patients with symptoms of hypothyroidism but thyroid function tests within the reference range: randomised double blind placebo controlled crossover trial. BMJ 323:891– 895 15. Bønaa KH, Arnesen E 1992 Association between heart rate and atherogenic blood lipid fractions in a population. The Tromsø Study. Circulation 86:394 – 405 16. Jorde R, Bønaa KH, Sundsfjord J 1999 Population based study on serum ionised calcium, serum parathyroid hormone, and blood pressure. The Tromsø study. Eur J Endocrinol 141:350 –357 17. Wechsler D 1987 Wechsler Memory Scale-Revised. Manual. New York: The Psychological Corp. 18. Reitan RM, Wolfson D 1993 The Halstead-Reitan Neuropsychological Test Battery. 2nd ed. Tucson, AZ: Neuropsychology Press 19. Golden CJ 1978 Stroop color and word test. Chicago: Stoelting 20. Lund-Johansen M, Hugdahl K, Wester K 1996 Cognitive function in patients with Parkinson⬘s disease undergoing stereotaxic thalamotomy. J Neurol Neurosurg Psychiatry 60:564 –571 21. Wechsler D 1955 Wechsler Adult Intelligence Scale. New York: The Psychological Corp. 22. Delis DC, Kramer JH, Kaplan E, Ober BA 1987 California Verbal Learning Test: Adult Version. San Antonio, TX: The Psychological Corp. 23. Spreen O, Strauss E 1998 A compendium of neuropsychological tests. 2nd ed. New York: Oxford University Press 24. Miller EN 1990 California Computerized Assessment Package. Los Angeles: Norland Software 25. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J 1961 An inventory for measuring depression. Arch Gen Psychiatry 4:225–236 26. Goldberg D, Williams P 1988 A user’s guide to the General Health Questionnaire. Windsor, UK: NFER-Nelson 27. Huppert FA, Walters DE, Day NE, Elliott BJ 1989 The factor structure of the General Health Questionnaire (GHQ-30). Br J Psychiatry 155:178 –185 28. Jaeschke R, Guyatt G, Cook D, Harper S, Gerstein HC 1994 Spectrum of quality of life impairment in hypothyroidism. Qual Life Res 3:323–327 29. Bjøro T, Holmen J, Kruger O, Midthjell K, Hunstad K, Schreiner T, Sandnes L, Brochmann H 2000 Prevalence of thyroid disease, thyroid dysfunction and thyroid peroxidase antibodies in a large, unselected population. The health study of Nord-Trøndelag (HUNT). Eur J Endocrinol 143:639 – 647 30. Hollowell JG, Staehling NW, Flanders WD, Hannon WH, Gunter EW, Spencer CA, Braverman LE 2002 Serum TSH, T4, and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab 87:489 – 499 31. Kudrajavcev T 1978 Neurologic complications of thyroid dysfunction. Adv Neurol 19:619 – 636 32. Leentjens AFG, Kappers EJ 1995 Persistent cognitive defects after corrected hypothyroidism. Psychopathology 28:235–237 33. Baldini IM, Vita A, Mauri MC, Amodei V, Carrisi M, Bravin S, Cantalamessa L 1997 Psychopathological and cognitive features in subclinical hypothyroidism. Prog Neuropsychopharmacol Biol Psychiatry 21:925–935 34. del Ser Quijano T, Delgado C, Martinez Espinosa S, Vazquez C 2000 Cognitive deficiency in mild hypothyroidism. Neurologia 15:193–198 35. Kirkegaard C, Faber J 1998 The role of thyroid hormones in depression. Eur J Endocrinol 138:1–9 36. Bono G, Fancellu R, Blandini F, Santoro G, Mauri M 2004 Cognitive and affective status in mild hypothyroidism and interactions with l-thyroxine treatment. Acta Neurol Scand 110:59 – 66 37. Pop VJ, Maartens LH, Leusink G, van Son MJ, Knottnerus AA, Ward AM, Metcalfe R, Weetman AP 1998 Are autoimmune thyroid dysfunction and depression related? J Clin Endocrinol Metab 83:3194 –3197 38. Zulewski H, Mu¨ller B, Exer P, Miserez AR, Staub JJ 1997 Estimation of tissue hypothyroidism by a new clinical score: evaluation of patients with various grades of hypothyroidism and controls. J Clin Endocrinol Metab 82:771–776 39. Canaris GJ, Steiner JF, Ridgway EC 1997 Do traditional symptoms of hypothyroidism correlate with biochemical disease? J Gen Intern Med 12:544 –550 40. Canaris GJ, Manowitz NR, Mayor G, Ridgway EC 2000 The Colorado thyroid disease prevalence study. Arch Intern Med 160:526 –534 41. Kapoor D, Jones TH 2005 Smoking and hormones in health and endocrine disorders. Eur J Endocrinol 152:491– 499 42. Biondi B, Palmieri EA, Lombardi G, Fazio S 2003 The importance of using strict inclusion criteria in subclinical hypothyroid studies. Am J Med 114:76

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.