Normal TSH levels: Still an elusive goal? Belgian Thyroid Club 36th meeting May 29th 2010

Bernard Corvilain Hôpital Erasme ULB

●

Serum TSH is the most sensitive index of thyroid function ► It’s important to determine a reference range or better a “normal” or “optimal” range for TSH

Reference intervals are not equivalent to optimal ranges ●

Reference intervals: Limits for a biological parameter between which 95% of the « normal » population is included. It only denotes what are usual values in the population, and does not directly correlate with the ranges for optimal health

Reference intervals Normal distribution of a biological value: bell shaped curve Exclusion of 2.5% of Individuals with the highest results

Exclusion of 2.5% of Individuals with the lowest results 95 % Reference range Lower limit

Upper limit

Reference intervals ≠ optimal ranges ●

Optimal range: Range for a biological parameter considered as optimal in terms of health or related complications and diseases –

The example of cholesterol

Optimal t-cholesterol < 190 mg/dl

optimal range Reference range

●

●

The difficulty in establishing TSH reference intervals is the choice of the « normal » euthyroid population: age, ethnicity, iodine intake, underlying medical condition and the choice of a reproducible TSH assay The difficulty in establishing TSH optimal range is to determine which level of TSH is associated with optimal health

TSH reference intervals ●

TSH reference intervals should be established from the 95% confidence limits of the log transformed values of at least 120 screened normal volunteers who have: –

No detectable thyroid autoantibodies

–

No personal or family history of thyroid disease

–

No visible or palpable goiter

–

No medications except estrogens

~0.3

C Spencer

TSH reference intervals ●

●

●

General agreement on TSH lower limit ~ 0.4 µU/ml A lot of discussion on TSH upper limit: 2.5 µU/ml ? 4 µU/ml? General agreement that specific ranges are needed according to age, ethnicity, iodine intake or underlying medical condition

TSH reference intervals Influence of age

TSH distribution by age groups in the United States: NHANES III study

TSH < 0.4:

TSH > 4.5:

2.7 % (20-29) 0.9% (50-59) 2.6% (80+)

2.0 % (20-29) 3.1 % (50-59) 12% (80+)

Surks, M. I. et al. J Clin Endocrinol Metab 2007;92:4575-4582

Copyright ©2007 The Endocrine Society

TSH reference intervals Influence of ethnicity

7.46

5.25

0.24

0.20

Boucai L, Surks MI. Reference limits of serum TSH and free T4 are significantly influenced by race and age in an urban outpatient medical practice. Clin Endocrinol (Oxf) 2009;70:788-93

TSH reference intervals Influence of iodine nutrition status

Influence of iodine status on TSH distribution > 65 yr

Jutland population (relatively low iodine intake) Iceland population (relatively high iodine intake)

Laurberg, P. et al. J Clin Endocrinol Metab 1998;83:765-769 Copyright ©1998 The Endocrine Society

TSH reference intervals Influence of underlying medical condition (Pregnancy)

TSH reference intervals during pregnancy 6

5

Upper reference 4.1 limit

4.80 4.16

4

By courtesy of Daniel Glinoër

3.67 Serum TSH (mIU/L)

3

2.93 2.51

2

Median

1

Lower reference 0.4 0.01 limit

2.15

1.03

1.23 0.53

1.35

0.12 0.03

0.09

0.31 0.20 0.02

First Trimester (n=7)

Second Trimester (n=5)

Third Trimester (n=2)

0.4

TSH reference intervals ●

●

TSH lower reference limit: ~ 0.4 µU/ml It is not possible to establish a universal TSH upper limit. It will depend upon the population being studied, the underlying pathology and iodine intake of that population. ~ 4 µU/ml ?

TSH normal (optimal) values

Is there a morbi-mortality associated with TSH value outside the reference range? ●

●

Is it harmful to be below the lower limit of TSH? Is it harmful to be above the upper limit of TSH ?

Cumulative Incidence of Atrial Fibrillation according to basal thyroid status

JAMA: Vol. 295 No. 9 2006

Subclinical hyperthyroidism and mortality

Subclinical Hypothyroidism and mortality

Is there a morbi-mortality associated with TSH value outside the reference range? ●

●

Is it harmful to be below the lower limit of TSH? YES Is it harmful to be above the upper limit of TSH ? No definitive data. Importance of age

Is there a morbidity associated with TSH values within the normal range? ●

Thyroid cancer (TSH growth factor)

●

Cholesterol

●

BMI

●

Osteoporosis

●

Atrial fibrillation

Thyroid function

TSH within the normal range and prevalence of malignancy in a thyroid nodule

Boelaert, K. et al. J Clin Endocrinol Metab 2006;91:4295-4301 Copyright ©2006 The Endocrine Society

TSH within the normal range and risk of malignancy in a thyroid nodule

TSH values within the normal range and Cholesterol (Hunt Study) ♂: 2.4% (1.7-3.2) change in non HDL cholesterol per 1 µU/ml TSH change

♀: 1.7% (1.2-2.3)

Asvold et al EJE 2007 156: 181-186

TSH values within the normal range and BMI

Knudsen, N. et al. J Clin Endocrinol Metab 2005;90:4019-4024

TSH values within the normal range and bone mineral density

Bone 2007 40: 1128-1134

TSH within the normal range and atrial fibrillation

High-Normal Thyroid Function and Risk of Atrial Fibrillation: The Rotterdam Study ●

Prospectively investigate the role of parameters of thyroid function in the development of Atria Fibrillation in euthyroid patients –

part of the population-based Rotterdam Study

–

1426 participants

–

Age > 55 yrs

–

TSH interval 0.4-4 µU/ml

–

median follow-up of 8 years Arch Intern Med. 2008;168(20):2219-2224

Serum TSH level and risk of atrial fibrillation (The Rotterdam study) 105 new cases of atrial fibrillation were identified

2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,2 0

P value for trend 0,02

Risque de FA (HR)

0,41,04

1,051,51

1,522,16

TSH value (µU/ml)

2,173,98

1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,2 0

Serum FT4 level and risk of atrial fibrillation (The Rotterdam study)P value for trend 0,06

Risque de FA (HR)

11-14,4

14,515,9

16-17,9 18-25

FT4 value (pmol/L)

●

The range of variation of thyroid hormone within an individual are within narrow limits by comparison with variation between individuals ► even values within the population reference range can be abnormal for an individual patient. Each individual has his own set point

Population reference range

Population reference range

Subclinical hypo ?

Subclinical hyper ?

Conclusions (1) ● ●

●

●

If we accept the data, there is no optimal value for TSH The increase in thyroid cancer risk could be due to trophic effect of TSH The influence of TSH on other parameters could be due to small differences in thyroid function i.e small differences in T3 nuclear receptor activation A normal low TSH level could have a favorable impact on cholesterol, weight and thyroid cancer, a negative impact on osteoporosis and atrial fibrillation and no impact on other factors

N Sarkosy

B Obama

Conclusions (2) ●

Several weaknesses in these studies –

Only one measurement of TSH at baseline

–

TSH value may be modified by intercurrent disease

–

Cross-sectional >> prospective studies

–

Relation of TSH with non independent variables (BMI, Cholesterol, Osteoporosis, age …)

–

A lot of correlations…. And correlation between two variables does not imply a causal connection

Theory of the stork

Storks bring babies….

Relation between the number of brooding storks in Germany and the number of new born babies between 1965 and 1980

Nature 332, 495 (7 April 1988) | A new parameter for sex education HELMUT SIES

Conclusions (3) ●

●

●

The upper TSH reference limit is not necessary the therapeutic threshold for initiating T4 replacement therapy particularly in old patients Interventional studies are needed if we want to establish a causal relationship between TSH within the normal ranges and other parameters If these results are explained by slight differences in T3 receptor activation, it underline the importance of clinical trials with selective T3 analogs to test in obesity or hypercholesterolemia or in the prevention of atrial fibrillation

Thank you for your attention !