The Effects of Levothyroxine Replacement or Suppressive Therapy on Health Status, Mood, and Cognition

Abstract

Context:

TSH-suppressive doses of levothyroxine (L-T4) have adverse effects on bone and cardiac function, but it is unclear whether central nervous system function is also affected.

Objective:

The aim of the study was to determine whether women receiving TSH-suppressive L-T4 doses have decrements in health status, mood, or cognitive function.

Design and Setting:

A cross-sectional comparison was made among three groups of women in an academic medical center research clinic.

Patients:

Twenty-four women receiving chronic TSH-suppressive L-T4 doses, 35 women receiving chronic replacement L-T4 doses, and 20 untreated control women participated in the study.

Interventions:

Subjects underwent testing at a single outpatient visit.

Main Outcome Measures:

We measured health status (SF-36), mood (Profile of Mood States, Symptom Checklist 90-R, Affective Lability Scale), and cognitive function (declarative memory [Paragraph Recall], working memory [N-back, Subject Ordered Pointing], motor learning [Pursuit Rotor, Motor Sequence Learning Test], and executive function [Letter Cancellation Test, Trail Making Test, Iowa Gambling Test]).

Results:

Women receiving TSH-suppressive or replacement L-T4 doses had decrements in health status and mood compared to healthy controls. These decrements were more pronounced in women receiving replacement, rather than suppressive, L-T4 doses. Memory and executive function were not affected in either treated group, compared to healthy controls.

Conclusions:

Women receiving TSH-suppressive doses of L-T4 do not have central nervous system dysfunction due to exogenous subclinical thyrotoxicosis, but TSH-suppressed and L-T4-replaced women have slight decrements in health status and mood that may be related to self-knowledge of the presence of a thyroid condition or other uncharacterized factors. These mood alterations do not impair cognitive function.

Levothyroxine (L-T4) doses that suppress serum TSH are used to prevent growth of thyroid nodules or cancer. However, L-T4 suppressive therapy has adverse effects, particularly on the heart and bones (1). L-T4 suppression therapy may also affect central nervous system function, but few studies have addressed this issue, and the results were inconsistent (2–6).

We investigated quality of life, mood, and cognition in women on L-T4 suppressive therapy, compared to women receiving L-T4 replacement doses or women with no history of thyroid disease. We employed sensitive measures of executive function and memory because these cognitive domains are affected in subjects with thyroid dysfunction (7) and animal studies support a major role for L-T4 in brain areas that mediate these functions (8). We hypothesized the following: 1) subjects receiving suppressive doses of L-T4 have decrements in health status and mood compared to healthy subjects and those receiving replacement doses of L-T4; and 2) subjects receiving suppressive doses of L-T4 have decrements in executive function and memory, which may be related to changes in health status or mood.

Subjects and Methods

Experimental subjects

Three groups of women were recruited. No subjects had acute or chronic illness or were on medications that affect thyroid hormone levels, mood, or cognition. Stable doses of oral contraceptives were allowed. Testing occurred during the first 14 days after onset of menstrual bleeding. Data on thyroid history, medical conditions, and medication use are provided in Supplemental Table 1 (published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org).

L-T4 suppressed group

In this group, 24 women (ages 22–53 y) were receiving L-T4 in doses that suppressed TSH to low or undetectable levels with normal or minimally elevated free T₄ (fT4) levels. Subjects were recruited from the authors' clinics and through review of abnormal TSH results from the clinical laboratory. Sixteen women had a history of low-risk thyroid cancer, whereas eight were inadvertently overtreated for hypothyroidism. None of these eight subjects had requested higher L-T4 doses or were aware that their doses were excessive. All subjects were diagnosed as adults, and all thyroid cancer patients had undergone total thyroidectomy and radioactive iodine ablation. Three subjects had received radioactive iodine for Graves' disease. Radioactive iodine had been administered at least 9 months before study (mean, 4 y; range, 9 mo to 10 y). Subjects had received L-T4 therapy for 1–25 years (mean, 6 y). L-T4 doses were stable for at least 3 months. Thyroid cancer patients had no evidence of disease on routine monitoring.

L-T4 euthyroid group

In this group, 35 women (ages 21–57 y) were receiving L-T4 for primary hypothyroidism (n = 29), hypothyroidism after I-131 therapy for Graves' disease (n = 3), postpartum thyroiditis leading to permanent hypothyroidism (n = 1), or thyroidectomy for nodular goiter or a distant history of low-risk thyroid cancer (n = 2). Subjects were recruited as part of a larger longitudinal study examining cognition in treated hypothyroid patients and were matched to the other two groups by age, estrogen status, and education. All subjects were diagnosed as adults and had received L-T4 therapy for 4 months to 35 years (mean, 11 y). Radioactive iodine had been administered to four subjects at least 5 years before the study (mean, 9 y; range, 5–13 y). All subjects had past elevated TSH levels. L-T4 doses were stable for at least 3 months.

Healthy control group

This group included 20 healthy women (ages 20–54 y) with normal TSH levels. Subjects were recruited by flyers and word of mouth from the general population.

Experimental design

The protocol was approved by the Oregon Health & Science University Institutional Review Board, and subjects gave written informed consent.

Screening visit

Subjects were screened for general health, medicines, thyroid status, and mood or cognitive disorders by history, physical examination, and laboratory testing. General intelligence was measured by the Wechsler Adult Intelligence Scale-Revised (WAIS-R) Vocabulary subtest (9).

Testing visit

Within 6 weeks, subjects returned for a 3- to 4-hour testing visit. Subjects refrained from taking L-T4 that morning. TSH, fT4, and free T₃ (fT3) levels were obtained. Subjects self-completed the following measures of health status and mood:

The SF-36 Health Survey (SF-36).

a questionnaire about general health (10). Higher scores reflect better health status and well-being.

The Profile of Mood States (POMS).

a questionnaire about mood (11). Higher scores reflect mood decrements, except for the vigor subscale, where higher scores reflect improved mood.

The Symptom Checklist 90-Revised (SCL90-R).

a questionnaire that evaluates psychological symptoms (12). Higher scores indicate increased psychological distress.

The Affective Lability Scale (ALS).

where subjects rate the tendency of their moods to shift between baseline and anger, depression, elation, or anxiety (13). Higher scores indicate increased lability of mood.

Cognitive tests were administered by a single experienced research assistant. Based on existent literature and our previous studies, we did not utilize a battery of general cognitive measures, but rather used sensitive measures targeted to specific domains likely to be affected by altered thyroid status.

Declarative memory

Paragraph Recall (verbal memory).

Subjects were read a brief story and verbally recalled it immediately and after 30 minutes. The outcome measure was the number of story elements recalled at each interval (14).

Working memory (also considered part of executive function)

N-Back Test.

A series of letters was presented one at a time on a computer screen. Each time a letter appeared, subjects responded by naming the letter they had seen on the previous screen (1-Back). The task was repeated with an increase in memory load by imposing intervening letters that the subjects had to hold in mind while they responded to letters that had appeared 2-back and then 3-back. Outcome measures were the total number correct on target and the total number incorrect off target for each condition (15).

Subject Ordered Pointing (SOP).

Subjects were presented with a series of computer screens with abstract drawings on them (6, 8, 10, or 12 per screen). Each screen in a set showed the same array of drawings, but in a different spatial arrangement. The subject was instructed to indicate one drawing per screen and to avoid choosing the same drawing on subsequent screens in the set. Subjects erred when they chose a drawing that had been previously chosen. Each set was repeated three times. The outcome measure was the number of errors across each screen set (16).

Motor learning

Pursuit Rotor.

Subjects held a photosensitive wand to maintain contact with a 2-cm light disk rotating on a turntable (model 30014; Lafayette Instrument Company). Two blocks of eight 20-second trials were administered, with a 20-second rest after each trial and a 60-second rest period after four trials. After a 30-minute interval, the two blocks were repeated. The outcome measure was the mean total time the stylus remained on target (17).

Motor Sequence Learning Test.

The subject memorized two keypress sequences on a computer, each associated with a letter of the alphabet. A “T” was associated with the sequence 1–3–2, and an “H” was associated with 3–1–2. When the “T” or “H” appeared on the screen, the subject performed the appropriate sequence as quickly as possible. Subjects performed 10 blocks of 18 trials each. The outcome was the average total movement time (time from character presentation to completion of the keypress sequence) (18).

Executive function

Attention/Concentration: The Letter Cancellation Test.

The subject was presented with a sheet of paper with six lines of 52 letters in random sequence. The subject was instructed to circle two specific letters each time they appeared, as quickly as possible. The score was the number of errors and the time taken to complete the task (19).

Trail Making Test.

The subject was instructed to connect circles as quickly as possible without lifting the pen from the paper. In part A, the circles were numbered, and the subject drew lines connecting the numbers in ascending order (1–2–3, etc). In part B, the circles included both numbers and letters; the subject drew lines to connect the circles in ascending order, alternating between numbers and letters (1-A-2-B-3-C, etc). Subjects were scored on the number of errors and time to complete the task (20).

Decision Making: The Iowa Gambling Test.

This task consisted of four face-down decks of cards on a computer screen. The subject chose cards from any deck, with each card resulting in the gain or loss of money. The subject was unaware that two decks were advantageous (small gains but smaller losses), whereas two were disadvantageous (large gains but even larger losses). The subject's choices were classified as advantageous (X) or disadvantageous (Y), with a net score of X-Y, over five trials of 100 cards each. This task assesses real-life decision making and responses to rewards and punishments (21).

Analytic methods

TSH was measured by immunochemiluminometric assay (Beckman Coulter); functional sensitivity, 0.02 mU/L; normal range, 0.34–5.60 mU/L; intraassay coefficient of variation (CV), 9.5% at 0.03 mU/L and 4.7% at 11.6 mU/L; interassay CV, 11% at 0.04 mU/L, 5% at 0.70 mU/L, and 5.8% at 24.94 mU/L. fT4 was measured by direct equilibrium dialysis (Quest Diagnostics); sensitivity, 0.08 ng/dL; normal range, 0.8–2.7 ng/dL; intraassay CV, 5.7% at 0.27 ng/dL and 1% at 4.6 ng/dL; interassay CV, 6.8% at 0.3 ng/dL and 1.6% at 3.8 ng/dL. fT3 was measured by tracer dialysis (Quest Diagnostics); sensitivity, 25 pg/dL; normal range, 210–440 pg/dL; intraassay CV, 6%; interassay CV, 4%.

Statistical methods

Subscales of each measure were analyzed together using linear repeated measures analyses (Proc Mixed, SAS/STAT; SAS Institute Inc) or, for binary outcomes, generalized estimating equations (Proc Genmod; SAS Institute Inc). These methodologies allow for correlation between subscale measures for each subject. For continuous outcomes, we analyzed the data using both a compound symmetric and a heterogeneous compound symmetric covariance structure, and we chose the final covariance structure to minimize the Bayes Information Criterion. These models included adjustments for age, WAIS-R, years in school, body mass index (BMI), and estrogen status. We allowed for different SD values within each group. Analysis of binary outcomes used compound symmetric covariance matrices and was unadjusted to the more limited nature of the data.

To limit the effect of multiple comparisons, an initial assessment of interaction between group and subscale was obtained for each set of subscales. If significant at level 0.10, a comparison of groups was conducted for each subscale. If the interaction was not significant, the comparison of groups was conducted for the set of scales as a whole (the main effect of group was analyzed, dropping the interaction from the model). Pairwise comparisons of groups, on the main effect or subscale level, were conducted only if overall evidence of a group effect was observed at level 0.05. We show results of pairwise comparisons with and without further multiple comparison adjustments using Holm's method (22). For outcomes where there were significant differences between groups at level 0.05, we examined relationships between the outcome and TSH, fT4, and fT3, using the same repeated measures methodology as for comparing groups, but substituting, in separate models, the selected hormone for the categorical group variable.

Results

Clinical parameters and thyroid function tests (Table 1)

Table 1.

Clinical Parameters and Thyroid Function Tests in the Three Groups

	Healthy Controls	L-T4 Euthyroid	L-T4 Suppressed	P Value
n	20	35	24
Age, y	40.0 ± 2.2	41.5 ± 1.7	38.9 ± 1.8	.58
WAIS-R	10.25 ± 0.62	10.49 ± 0.30	10.17 ± 0.45	.82
Years in school	16.1 ± 0.5	16.7 ± 0.5	15.9 ± 0.6	.69
Estrogen status	60% Pre-none, 20% Pre-on, 20% Post-none	60% Pre-none, 17% Pre-on, 23% Post-none	67% Pre-none, 17% Pre-on, 17% Post-none	.98
BMI, kg/m2	29.1 ± 1.7	28.1 ± 0.8	25.6 ± 0.9	.09
L-T4 dose, μg/kg	NA	1.46 ± 0.09	1.93 ± 0.11c	.001
TSH, mU/L	2.16 ± 0.20	1.83 ± 0.16	0.13 ± 0.02a,c	<.0001
fT4, ng/dL	1.38 ± 0.06	1.61 ± 0.06a	2.26 ± 0.09a,c	<.0001
fT3, pg/dL	243 ± 9	217 ± 6b	239 ± 7d	.02

Abbreviations: Pre-none, premenopausal, no hormone treatment; Pre-on, premenopausal on hormone treatment; Post-none, postmenopausal, no hormone treatment; NA, not applicable. Data are expressed as mean ± SEM, unless otherwise specified.

^aSignificantly different than healthy control group at level 0.01 by post hoc tests.

^bSignificantly different than healthy control group at level 0.05 by post hoc tests.

^cSignificantly different than L-T4 euthyroid group at level 0.01 by post hoc tests.

^dSignificantly different than L-T4 euthyroid group at level 0.05 by post hoc tests.

The groups were well matched for age, WAIS-R, years in school, and estrogen status, and hence unadjusted and adjusted analyses are similar. BMI was lower in the L-T4 suppressed group, although this did not reach statistical significance. As expected, L-T4 doses were higher in L-T4 suppressed subjects compared to L-T4 euthyroid subjects (1.46 ± 0.09 vs 1.93 ± 0.11 μg/kg/d; P = .001). Mean TSH levels were similar in the healthy control and L-T4 euthyroid groups (2.16 ± 0.20 and 1.83 ± 0.16 mU/L, respectively) but, as expected, were lower in the L-T4 suppressed group (0.13 ± 0.02 mU/L; P < .0001). Within the L-T4 suppressed group, TSH levels were < 0.1 mU/L in 10 subjects and 0.1–0.33 mU/L in 14 subjects. Compared to the healthy control group, mean fT4 levels were higher in the L-T4 euthyroid group and were even higher in the L-T4 suppressed group (1.38 ± 0.06, 1.61 ± 0.06, and 2.26 ± 0.09 ng/dL, respectively; P < .0001). Three L-T4 suppressed subjects had slightly elevated fT4 levels, whereas the other 21 subjects had normal fT4 levels. Mean fT3 levels were lower in the L-T4 euthyroid group, compared to the other two groups (217 ± 6, 243 ± 9, and 239 ± 7 pg/dL, respectively; P = .02). fT3 levels were slightly low in four healthy control subjects, 17 L-T4 euthyroid subjects, and four L-T4 suppressed subjects. No subjects had elevated fT3 levels. Individual thyroid hormone levels are shown in Supplemental Table 2.

Health status and mood (SF-36, POMS, SCL90-R, ALS) (Table 2)

Table 2.

Health Status and Mood Measures in the Three Groups

Measure	Healthy Controls	L-T4 Euthyroid	L-T4 Suppressed	P Values for Comparing Groups
n	20	35	24
SF-36
MCS (mental component summary)	50.9 ± 0.8	44.7 ± 1.4a	46.3 ± 1.1a	.0001
PCS (physical component summary)	50.3 ± 1.8	51.7 ± 1.3	54.6 ± 0.9c	.04
GH (general health)	77.9 ± 3.2	74.3 ± 3.1	74.6 ± 2.9	.67
MH (mental health)	63.8 ± 1.8	56.3 ± 1.9b	57.6 ± 1.6c	.008
VT (vitality)	66.7 ± 3.9	52.1 ± 3.1a	59.8 ± 3.9	.009
BP (bodily pain)*	28% high	29% high	52% high	.27
PF (physical functioning)*	50% high	49% high	48% high	.27
RP (role physical)*	72% high	71% high	87% high	.27
SF (social functioning)*	72% high	49% high	70% high	.27
RE (role emotional)*	100% high	83% high	83% high	.27
POMS**
A (anger)	2.8 ± 0.4	4.2 ± 0.8	3.0 ± 0.5	.54
C (confusion)	5.4 ± 0.5	6.3 ± 0.5	6.1 ± 0.6	.48
D (depression)	3.7 ± 1.0	5.1 ± 1.2	3.5 ± 0.9	.45
F (fatigue)	5.5 ± 1.0	8.7 ± 1.0c	5.8 ± 0.9d	.02
T (tension)	6.7 ± 0.8	6.1 ± 0.6	7.0 ± 0.8	.66
V (vigor)	17.4 ± 1.3	14.3 ± 0.9	16.7 ± 1.3	.06
SCL90-R
GSI (global severity index)	44.2 ± 2.0	50.4 ± 1.5b	49.8 ± 2.1	.03
PST (positive symptoms total)	45.3 ± 1.9	49.9 ± 1.5	50.0 ± 2.1	.12
PSDI (positive symptom distress index)	47.2 ± 1.9	50.9 ± 1.3	47.7 ± 1.7	.12
ANX (anxiety)	44.0 ± 1.8	44.9 ± 1.5	48.5 ± 1.9	.28
DEP (depression)	46.0 ± 1.7	51.8 ± 1.5b	49.2 ± 1.8	.04
HOS (hostility)	45.3 ± 1.4	48.2 ± 1.2	47.5 ± 1.3	.39
IS (interpersonal sensitivity)	48.3 ± 1.6	52.8 ± 1.5	52.4 ± 1.6	.14
OC (obsessive-compulsive)	47.3 ± 2.1	53.9 ± 1.4b	53.2 ± 2.1c	.02
SOM (somatization)	47.3 ± 1.6	51.7 ± 1.6	49.6 ± 2.3	.13
Phob (phobic anxiety)***	70% low	74% low	75% low	.94
Par (paranoid ideation)***	60% low	60% low	63% low	.94
Psy (psychoticism)***	70% low	69% low	54% low	.94
ALS
Depression	0.7 ± 0.1	1.2 ± 0.1a	0.9 ± 0.1	.006
Anger****	50% score > 0	60% score > 0	63% score > 0	.68
Anxiety****	60% score > 0	80% score > 0	88% score > 0	.13
Bipolar****	75% score > o	94% score > 0	88% score > 0	.18
Elation****	75% score > 0	94% score > 0	88% score > 0	.18

Data are expressed as mean ± SEM, unless otherwise specified. P values for continuous outcomes were adjusted for age, years of education, WAIS-R, BMI, and estrogen status. Boldface indicates measures that were statistically different among groups.

^*For PF, RP, RE, SF, and BP, the distributions within each group were highly skewed. We used the highest observed values of the skewed SF-36 subscales as the cut-points for producing a dichotomous measure. For BP, the highest observed value was 90, whereas for the other scales, the highest observed value was 100. RE was analyzed separately due to 100% scoring low in healthy control group.

^**POMS values were natural log-transformed before analysis because the raw data were skewed. The vigor subscale was analyzed separately because it was the only scale for which higher values represented improved mood.

^***These scores were calculated as high: > 44 for Phob, Psy, and High > 41 for Par, due to floor effects.

^****These scores were compared as the proportion positive between the groups because the measures were skewed and contained a large proportion of zeros.

^aSignificantly different from control at level 0.01 before and after multiple comparison adjustment.

^bSignificantly different from control at level 0.01 before multiple comparison adjustment and at level 0.05 after multiple comparison adjustment.

^cSignificantly different from control group at level 0.05 before multiple comparison adjustment, but not after adjustment.

^dSignificantly different from LT4 euthyroid group at level 0.05 before multiple comparison adjustment, but not after adjustment.

Compared to the healthy control group, the L-T4 euthyroid and L-T4 suppressed groups scored worse on the SF-36 mental component summary. Both groups also scored worse on the SF-36 mental health and SCL90-R obsessive-compulsive subscales, but these were no longer significant in the L-T4 suppressed group after adjustment for multiple comparisons. The L-T4 euthyroid group scored worse than the other groups on the SCL90-R global severity index, as well as the SF-36 vitality, the ALS and SCL90-R depression, and the POMS fatigue subscales; the latter was no longer significant after adjustment. There was only one subscale where the L-T4 suppressed group differed from the healthy control and euthyroid groups, with a better score on the SF-36 physical component summary, and this was no longer significant after adjustment. There were no effects on other health status or mood subscales.

Cognitive tests (Table 3)

Table 3.

Cognitive Measures in the Three Groups

Test	Healthy Controls	L-T4 Euthyroid	L-T4 Suppressed	P Values for Comparing Groups
Declarative memory
Paragraph recall
Immediate	13.5 ± 0.9	13.7 ± 0.6	13.9 ± 0.8	.81
30-min delay	11.8 ± 1.1	11.9 ± 0.7	12.2 ± 0.9	.81
Working memory
N-Back number correct on target
1-Back*	65%	85%a	58%b	.02
2-Back*	35%	59%a	46%b
3-Back	10.2 ± 0.4	10.7 ± 0.4	10.0 ± 0.6	.29
N-Back number incorrect nontarget
1-Back*	20%	21%	25%	.50
2-Back*	35%	21%	38%
3-Back	3.0 ± 0.3	2.9 ± 0.3	3.9 ± 0.5	.09
SOP errors
6	0.7 ± 0.1	0.5 ± 0.1	0.5 ± 0.1	.16
8	1.4 ± 0.2	0.9 ± 0.1	1.0 ± 0.1	.16
10	1.3 ± 0.2	1.2 ± 0.1	1.2 ± 0.2	.16
12	1.4 ± 0.2	1.6 ± 0.2	1.5 ± 0.2	.16
Motor learning
Pursuit Rotor Trial
Time on target, s
1	34.0 ± 2.5	33.2 ± 2.0	31.9 ± 2.2	.87
2	38.1 ± 2.5	36.5 ± 2.0	37.8 ± 2.4	.87
3	39.2 ± 2.6	38.5 ± 2.0	38.9 ± 2.6	.87
4	39.6 ± 2.4	39.0 ± 1.7	40.7 ± 2.4	.87
Motor Sequence Learning Test
Total movement time, s	1069 ± 48	1001 ± 43	1000 ± 43	.74
Executive function
Letter Cancellation Test
Time, s	96 ± 4	95 ± 3	100 ± 4	.54
Errors	2.6 ± 0.7	2.9 ± 0.7	2.5 ± 0.6	.95
% with no errors	20%	23%	21%	1.00
Trail Making Test
Time, s	25 ± 2	22 ± 1	21 ± 1
ABC time, s	57 ± 4	53 ± 3	61 ± 4	.30
% with errors	10%	11%	4%
% with ABC errors	25%	23%	17%	.63
Iowa Gambling Test
Net-1	1.3 ± 2.8	−4.6 ± 1.5	−3.9 ± 1.8	.88
Net-2	2.7 ± 2.3	6.5 ± 1.4	3.3 ± 1.8	.88
Net-3	7.9 ± 2.1	9.9 ± 1.4	8.8 ± 1.6	.88
Net-4	7.7 ± 2.1	10.1 ± 1.8	11.9 ± 1.7	.88
Net-5	9.7 ± 2.1	11.1 ± 1.8	9.3 ± 2.0	.88

Data are expressed as mean ± SEM, unless otherwise specified. Individual tests are grouped by cognitive subdomains (first column). P values for continuous outcomes were adjusted for age, years of education, WAIS-R, BMI, and estrogen status. Boldface indicates measures that were statistically different among groups.

^*These values were calculated as proportion of subjects for which each measure was ≥ 15 (for Correct on Target) or > 0 (for Incorrect Off Target) because there were floor effects.

^aSignificantly different than healthy control group at level 0.05 by post hoc tests.

^bSignificantly different than L-T4 euthyroid group at level 0.05 by post hoc tests.

Declarative memory

There were no differences among the groups in paragraph recall.

Working memory

The L-T4 euthyroid group performed slightly better on the 1-Back and 2-Back number correct on target, compared to the other two groups. There were no differences among the groups on the SOP test.

Motor learning

There were no differences among the groups in the Pursuit Rotor or Motor Sequence Learning tests.

Executive function

There were no differences among the groups in the Letter Cancellation, Trail Making, or Iowa Gambling tests.

Correlations between thyroid hormone levels and health status, mood, and cognition (Table 4)

Table 4.

Correlations Between Thyroid Hormone Levels and Significant Health Status, Mood, and Cognitive Measures in the Three Groups

Measure	fT4		fT3		TSH
	Coefficient	P Value	Coefficient	P Value	Coefficient	P Value
SF-36 mental component summary*	−2.93	.02	0.04	.01	0.32	.70
SF-36 physical component summary*	3.41	.006	0.01	.64	−0.15	.90
SF-36 mental health subscale*	−3.52	.08	0.06	.03	0.75	.60
SF-36 vitality subscale*	1.82	.66	0.12	.03	−0.90	.73
POMS fatigue subscale*	−0.29	.10	− 0.003	.20	0.16	.13
SCL90-R global severity index*	0.84	.67	−0.04	.11	1.99	.12
SCL90-R depression subscale*	−0.42	.83	−0.05	.07	0.72	.57
SCL90-R obsessive-compulsive subscale*	1.75	.38	− 0.008	.76	2.26	.08
ALS depression subscale*	0.02	.78	−0.002	.07	0.03	.59
N-Back number correct on target**	−3%	.86	−0.3%	<.0001	0.1%	.99

Correlations were analyzed by repeated measures methodology for outcomes that were different among the three groups, using separate models for each hormone. Significant coefficients are shown in bold with corresponding P values. A positive coefficient indicates that the measure increased with increasing hormone levels, whereas a negative coefficient indicates that the measure decreased with increasing hormone levels. The L-T4 suppressed group was excluded from the TSH correlation calculations due to multiple undetectable TSH levels and a skewed TSH distribution. Boldface indicates measures that were statistically different among groups.

^*For the health status and mood measures, the magnitude of the coefficient indicates the estimated change in the measure with a one-unit increase in the hormone level.

^**For the N-Back measure, coefficients were transformed to estimate the probability of obtaining a high score (15 or higher). The transformed coefficients are estimates of the risk ratios associated with a one-unit increase in the hormone level.

fT4 levels were inversely related to the SF-36 mental component and directly related to the SF-36 physical component summaries. fT3 levels were directly related to the SF-36 mental component summary and the mental health and vitality subscales, and were inversely related to N-Back number correct. However, in each case the magnitude of the relationship was small. TSH was not related to any of the outcomes.

Discussion

We investigated the effects of L-T4 suppressive therapy on health status, mood, and cognition in women with benign thyroid conditions or low-risk thyroid cancer. We utilized sensitive, validated measures for mood and cognitive domains likely to be affected by altered thyroid status. We included two matched control groups—healthy euthyroid women with no L-T4 therapy, and hypothyroid women receiving L-T4 with normal TSH levels.

We hypothesized that L-T4 suppressed women would have decrements in health status and/or mood. We did find slight decrements in SF-36 mental health scales and the SCL90-R obsessive-compulsive subscale in this group, although these diminished with adjustments for multiple comparisons. Importantly, the L-T4 euthyroid group had similar decrements in these scales, as well as additional decrements in the SCL90-R global severity index and depression subscale, and the SF-36 vitality, POMS fatigue, and ALS depression subscales. The only measure where the L-T4 suppressed subjects differed from control groups was a slightly better score on the SF-36 physical component summary, and this finding did not remain after adjustment. When considered as continuous variables, there were inconsistent correlations between fT4 or fT3 levels and some SF-36 measures, but these were unlikely to be clinically relevant, given their small magnitude.

Past studies reported decreased (4, 6) or unchanged (5) quality of life in subjects receiving L-T4 suppressive therapy. One study showed less depression in these patients (5). Additional studies reported decrements in health status or mood in endogenous subclinical thyrotoxicosis (23–27). However, these studies all lacked control groups with known thyroid disease. Larger, population-based studies did not find decrements in mood due to endogenous subclinical thyrotoxicosis (28–31). Of interest, studies consistently report that subjects with known thyroid disease have increased fatigue and decreased mood regardless of thyroid status (32–34).

These observational studies are supported by two blinded, placebo-controlled studies that altered L-T4 doses in L-T4-treated subjects. When mean TSH levels were lowered from 3.02 to 0.08 mU/L, subjects experienced no changes in health-related quality of life (35). In the second study, mood improved slightly when mean TSH levels were lowered from 2.15 to 0.17 mU/L (2).

Together with our findings, this literature shows that neither exogenous nor endogenous subclinical thyrotoxicosis is associated with decrements in health status or mood. Our findings highlight the importance of including a control group matched for the presence of thyroid disease. Without this, one could conclude that subjects receiving L-T4 suppression therapy have decrements in health status and mood due to an excessive dose of L-T4.

We found decrements in health status and mood in L-T4 euthyroid subjects, across a number of measures. Previous studies showed similar decrements in euthyroid subjects receiving L-T4 (32, 33, 36, 37). This has sometimes been attributed to inadequate normalization of thyroid hormone levels. In our study, mean TSH levels were similar in the healthy control and L-T4 euthyroid groups, although the L-T4 euthyroid subjects had lower mean fT3 levels. This could contribute to our findings, although L-T3 treatment studies have not consistently shown improvement in these measures (38). This also would not explain why our L-T4 suppressed subjects had similar decrements. Finally, a double-blinded study where L-T4 doses were altered to different TSH levels within the normal range did not show effects on psychological function (35). An alternative explanation is that self-knowledge of a thyroid diagnosis confers a lower health status and mood independently of current thyroid status, as discussed above.

Our second hypothesis was that L-T4 suppressed subjects would have decrements in memory or executive function, but we found no changes in sensitive measures of these cognitive domains. Decrements in executive function, psychomotor speed, and attention were found in one similar study of thyroid cancer patients receiving suppressive L-T4 doses (3). Their subjects had more medical and psychiatric comorbidities and were on higher L-T4 doses than our subjects. Large, population-based studies failed to find cognitive decrements in subjects with endogenous subclinical thyrotoxicosis (28, 29). Additionally, our results are concordant with the two blinded intervention studies described above (2, 35), which showed no decrements in cognitive function when L-T4 doses were increased to TSH-suppressed levels. These two studies induced short-term exogenous subclinical hyperthyroidism, whereas our current study investigated long-term use of suppressive doses of L-T4. Taken together, these studies suggest that neither exogenous nor endogenous subclinical thyrotoxicosis impairs memory or executive function.

Finally, despite decrements in mood, our L-T4 euthyroid group had intact measures of memory and executive function. This is concordant with past studies (39, 40), although one study reported decreased attention and verbal memory in euthyroid subjects receiving L-T4 (36). This study did not include a control group, and subjects were aware of their diagnoses. In contrast, when L-T4 doses were increased in a blinded fashion in euthyroid L-T4 treated subjects, lowering mean TSH levels from 3.02 to 0.83, there were no changes in cognitive function (35). We conclude that subjects on L-T4 with normal TSH levels do not have impairments in memory or executive function.

There are a number of strengths to our study: two well-matched control groups, and the use of sensitive, validated measures of health status, mood, and cognitive functions. The L-T4 treated euthyroid group allowed us to distinguish effects of the L-T4 dose from general effects due to the presence of a thyroid diagnosis. However, there are also limitations to our study: the numbers of subjects were relatively small for the number of comparisons we conducted. We performed group analyses together, progressing to pairwise comparisons only in the presence of significant group effects and examining subscales only when there was evidence of interaction. Our results for mood were consistent across different surveys, suggesting that our results are not an artifact of multiple measures. For most of the measures, P values were below .01, attesting to the strength of our observed associations. The groups were heterogeneous in terms of length of time of L-T4 therapy overall and at current doses, degree of TSH suppression in the L-T4 suppressed group, underlying cause and degree of hypothyroidism, and administration of radioactive iodine. The L-T4 replaced group had a longer mean duration of disease than the L-T4 suppressed group, which could have contributed to decreased mood. Ideally, these variables would be more tightly clustered, but this is not feasible in terms of recruitment. There were additional variables that could have affected results, including marital and socioeconomic status. Although we assume those variables were evenly distributed among the groups, we do not have data on them. BMI was lower in the L-T4 suppressed group, although this did not reach statistical significance, and we adjusted for this in our analysis. The minimum duration of a stable dose of L-T4 was 3 months, raising the question of whether this was sufficient time to observe effects on mood and cognition. Our previous studies and published reports showed changes in these parameters within 12 weeks of altering thyroid status (2, 4). Therefore, we believe our results accurately reflect subjects' thyroid status. Finally, the study was cross-sectional in design, so we could not ascertain longitudinal effects or causality of our observed mood alterations.

In summary, we found slight decrements in health status and mood in subjects treated with replacement or suppressive doses of L-T4, particularly greater fatigue and depression. This suggests that the known presence of thyroid disease has a negative effect on patients' perceived health status and mood. In contrast, we found little evidence that L-T4 therapy, as replacement or TSH suppression, alters memory or executive function. Patients receiving L-T4 who have normal or low TSH levels and significant affective or cognitive symptoms should be evaluated for other diagnoses.