NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease

Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.

C. Thyrotropin/Thyroid Stimulating Hormone (TSH)
For more than twenty-five years, TSH methods have been able to detect the TSH elevations that are
characteristic of primary hypothyroidism. Modern-day TSH methods however, with their enhanced sensitivity
are also capable of detecting the low TSH values typical of hyperthyroidism. These new methods are often
based on non-isotopic immunometric assay (M!" principles and are available on a variety of automated
immunoassay analy#er platforms. Most of the current methods are capable of achieving a functional sensitivity
of $.$%m&'( or less, which is a necessary detection limit for the full range of TSH values observed between
hypo- and hyperthyroidism. )ith this level of sensitivity, it is possible to distinguish the profound TSH
suppression typical of severe *raves+ thyroto,icosis (TSH - $.$. m&'(" from the TSH suppression ($.$. / $..
m&'(" observed with mild (subclinical" hyperthyroidism and in some patients with a non-thyroidal illness
(0T".
n the last decade the diagnostic strategy for using TSH measurements has changed as a result of the sensitivity
improvements in these assays. t is now recogni#ed that the TSH measurement is a more sensitive test than FT1
for detecting both hypo- and hyperthyroidism. !s a result, some countries now promote a TSH-first strategy for
diagnosing thyroid dysfunction in ambulatory patients (provided that the TSH method has a functional
sensitivity 2 $.$% m&'(". 3ther countries still favor the TSH 4 FT1 panel approach, because the TSH-first
strategy can miss patients with central hypothyroidism 5Section-6 71(f"8 or TSH-secreting pituitary tumors
5Section-6 71(g"i8 (19,195-197). !n additional disadvantage of the TSH-centered strategy is that the TSH-FT1
relationship cannot be used as a 9sanity chec:; for interferences or detection of unusual conditions
characteri#ed by discordance in the ratio of TSH'FT1 (Table .".
1. Specificity
(a" TSH Heterogeneity
TSH is a heterogeneous molecule and different TSH isoforms circulate in the blood and are present in the
pituitary e,tracts used for assay standardi#ation (Medical <esearch 7ouncil (M<7" =$'>>=". n the future,
recombinant human TSH (rhTSH" preparations might be used as primary standards for standardi#ing TSH
immunoassays (198). 7urrent TSH M! methods use TSH monoclonal antibodies that virtually eliminate cross-
reactivity with other glycoprotein hormones. These methods however, may detect different epitopes of
abnormal TSH isoforms secreted by some euthyroid individuals, as well as some patients with abnormal
pituitary conditions. For e,ample, patients with central hypothyroidism caused by pituitary or hypothalamic
dysfunction, secrete TSH isoforms with abnormal amounts of glycosylation and reduced biological activity.
These isoforms are measured as parado,ically normal or even elevated serum TSH concentrations by most
methods (195,197,199). (i:ewise, parado,ically normal serum TSH levels may be seen in patients with
hyperthyroidism due to a TSH-secreting pituitary tumor, that appears to secrete TSH isoforms with enhanced
biologic activity (196,200,201).
(b" Technical ?roblems
Technical problems, especially with the washing step, may result in falsely high TSH values (202). !dditionally,
any interfering substance in the specimen (eg heterophilic antibodies, H!M!" that produces a high bac:ground
or a false bridge between the capture and signal antibodies will create a high signal on the solid support that will
be read out as a falsely high result 5see Section-%768 (203,202).
(c" Methods for @etecting nterference with a TSH <esult
The conventional laboratory approach to verifying an analyte concentration such as dilution may not always
detect an interference problem. Since methods vary in their susceptibility to most interfering substances, the
most practical way to test for interference is to measure the TSH concentration in the specimen using a different
manufacturer+s method, and to chec: for a significant discordance between the TSH values. )hen the
- 31 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
variability of TSH measurements made on the same specimen with different methods e,ceeds e,pectations
(A>$B difference", interference may be present. Ciologic chec:s may also be useful to verify an une,pected
result. nappropriately low TSH values could be chec:ed by a T<H-stimulation test, which is e,pected to
elevate TSH more than %-fold (D1.$ m&'( increment" in normal individuals (204). n cases where TSH appears
inappropriately elevated, a thyroid hormone suppression test (.mg (-T1 or %$$Eg (-T6, po" would be e,pected
to suppress serum TSH more than F$ B by 1= hours in normal individuals.

Guideline 18. Investigation of Discordant Serum TSH Values in Ambulatory atients
A discordant TSH result in an a!ulator" #atient $it% sta!le t%"roid status a" !e a tec%nical error& S#eci'icit"
loss can result 'ro la!orator" error, inter'erin( su!stances (i&e& %etero#%ilic anti!odies), or t%e #resence o' an
unusual TSH iso'or (see )uideline 7 and Ta!le 1)& *%"sicians can re+uest t%at t%eir la!orator" #er'or t%e
'ollo$in( c%ec,s-
7onfirm specimen identity (i.e. have laboratory chec: for a switched specimen in the run".
)hen TSH is une,pectedly high, as: the laboratory to re-measure the specimen diluted, preferably in
thyroto,ic serum, to chec: for parallelism.
<eGuest that the laboratory analy#e the specimen by a different manufacturer+s method (send to a different
laboratory if necessary". f the between-method variability for a sample is A >$B, an interfering substance
may be present.
3nce a technical problem has been e,cluded, biologic chec:s may be usefulH
- &se a T<H stimulation test for investigating a discordant low TSH result, e,pect a %-fold
(D1.$ m&'( increment" response in TSH in normal individuals.
- &se a thyroid hormone suppression test to verify a discordant high TSH level. 0ormal
response to .mg of (-T1 or %$$Eg (-T6 administered p.o. is a suppressed serum TSH of
more than F$ B by 1= hours.
!. Sensitivity
Historically, the IGualityI of a serum TSH method has been determined from a clinical benchmar: - the assayJs
ability to discriminate euthyroid levels (K $.1 to 1.$ m&'(" from the profoundly low (-$.$. m&'(" TSH
concentration typical of overt *ravesJ thyroto,icosis. Most TSH methods now claim a detection limit of $.$%
m&'( or less (9third generation; assays" (202).
Guideline 1". Definition of #unctional Sensitivity
.unctional Sensiti/it" s%ould !e used to deterine t%e 0o$est 1etection 0iit o' t%e assa"&
TSH assay functional sensitivity is defined as a %$ B between-run coefficient of variation (7L" determined
by the recommended protocol (see *uideline %$".
TSH assay functional sensitivity is defined by the %$ B between-run coefficient of variation (7L" determined
by the recommended protocol (see *uideline %$".
- 32 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
Guideline !$. rotocol for TSH #unctional Sensitivity % &et'een()un recision
2easure %uan seru #ools co/erin( t%e assa" ran(e in at least 10 di''erent runs& T%e lo$est #ool /alue
s%ould !e 103 a!o/e t%e detection liit and t%e %i(%est #ool /alue s%ould !e 903 o' t%e
u##er assa" liit&
7arry-over should be assessed by analy#ing the highest pool followed by the lowest pool.
&se the same test mode as for patient specimens (i.e. singlicate or duplicate".
The instrument operator should be blinded to the presence of test pools in the run.
<uns should be spaced over a clinically representative interval (i.e. M to = wee:s for TSH in an outpatient
setting".
&se at least two different lots of reagents and two different instrument calibrations during the testing
period.
)hen running the same assay on two similar instruments, blind duplicates should be run on each
instrument periodically to verify correlation.
Manufacturers have largely abandoned the use of the 9analytical sensitivity; parameter for determining the
sensitivity of a TSH assay because it is calculated from the within-run precision of the #ero calibrator which
does not reflect the sensitivity of the test in clinical practice (126,127). nstead, a 9functional sensitivity;
parameter has been adopted (202). Functional sensitivity is calculated from the %$B between-run coefficient of
variation (7L" for the method and is used to establish the lowest reportable limit for the test (202).
Functional sensitivity should be determined by strictly adhering to the recommended protocol (*uideline %$"
that is designed to assess the minimum detection limit of the assay in clinical practice and ensure that the
parameter realistically represents the lowest detection limit of the assay. The protocol is designed to ta:e into
account a variety of factors that can influence TSH method imprecision in clinical practice. These includeH
Matri, differences between patient serum and the standard diluent
Nrosion of precision over time
(ot-to-lot variability in the reagents supplied by the manufacturer
@ifferences between instrument calibration and technical operators
7arry-over from high to low specimens (205)
The use of the functional sensitivity limit as the lowest detection limit is a conservative approach to ensure that
any TSH result reported is not merely assay 9noise;. Further, the %$B between-run 7L appro,imates the
ma,imum imprecision reGuired for diagnostic testing (Table >".
Guideline !1. #or *aboratories erforming TSH Testing
.unctional sensiti/it" is t%e ost i#ortant #er'orance criterion t%at s%ould in'luence t%e selection o' a TSH
et%od& *ractical 'actors suc% as instruentation, incu!ation tie, cost, and tec%nical su##ort t%ou(%
i#ortant, s%ould !e secondar" considerations& 0a!oratories s%ould use cali!ration inter/als t%at o#tii4e
'unctional sensiti/it", e/en i' re-cali!ration needs to !e ore 're+uent t%an recoended !" t%e anu'acturer-
Select a TSH method that has a functional sensitivity 2 $.$% m&'(
Nstablish functional sensitivity independent of the manufacturer by using *uideline %$
There is no scientific Oustification to refle, from a less sensitive to a more sensitive test. (nsensitivity
causes falsely high, not falsely low, values that are missed by refle, testingP"
+. TSH )eference Intervals
@espite some gender, age and ethnicity-related differences in TSH levels revealed by the recently published
0H!0NS &S survey, it is not considered necessary to adOust the reference interval for these factors in
clinical practice (18). Serum TSH levels e,hibit a diurnal variation with the pea: occurring during the night and
- 33 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
the nadir, which appro,imates to >$B of the pea: value, occurring between .$$$ and .M$$ hours (123,124).
This biologic variation does not influence the diagnostic interpretation of the test result since most clinical TSH
measurements are performed on ambulatory patients between $=$$ and .=$$ hours and TSH reference intervals
are more commonly established from specimens collected during this time period. Serum TSH reference
intervals should be established using specimens from T?3!b-negative, ambulatory, euthyroid subOects who
have no personal or family history of thyroid dysfunction and no visible goiter. The variation in the reference
intervals for different methods reflects differences in antibody epitope recognition by the different :it reagents
and the rigor applied to the selection of appropriate normal subOects.
Serum TSH concentrations determined in normal euthyroid subOects are s:ewed with a relatively long ItailI
towards the higher values of the distribution. The values become more normally distributed when log-
transformed. For reference range calculations, it is customary to log-transform the TSH results to calculate the
F>B reference interval (typical population mean value K..> m&'(, range $.1 to 1.$ m&'( in iodide-sufficient
populations" (202,206). However, given the high prevalence of mild (subclinical" hypothyroidism in the general
population, it is li:ely that the current upper limit of the population reference range is s:ewed by the inclusion
of persons with occult thyroid dysfunction (18)&
Guideline !!. TSH )eference Intervals
TSH reference intervals should be established from the 95 % confidence limits of the log-transformed values
of at least 120 rigorously screened normal euthyroid volunteers who have
0o detectable thyroid autoantibodies, T?3!b or Tg!b (measured by sensitive immunoassay"
0o personal or family history of thyroid dysfunction
0o visible or palpable goiter
0o medications (e,cept estrogen".
(a" TSH &pper <eference (imits
3ver the last two decades, the upper reference limit for TSH has steadily declined from K.$ to appro,imately
K1.$-1.> m&'(. This decrease reflects a number of factors including the improved sensitivity and specificity of
current monoclonal antibody based immunometric assays, the recognition that normal TSH values are log-
distributed and importantly, improvements in the sensitivity and specificity of the thyroid antibody tests that are
used to pre-screen subOects. The recent follow-up study of the )hic:ham cohort has found that individuals with
a serum TSH A%.$ m&'( at their primary evaluation had an increased odds ratio of developing hypothyroidism
over the ne,t %$ years, especially if thyroid antibodies were elevated (35). !n increased odds-ratio for
hypothyroidism was even seen in antibody-negative subOects. t is li:ely that such subOects had low levels of
thyroid antibodies that could not be detected by the insensitive microsomal antibody agglutination tests used in
the initial study (207). Nven the current sensitive T?3!b immunoassays may not identify all individuals with
occult thyroid insufficiency. n the future, it is li:ely that the upper limit of the serum TSH euthyroid reference
range will be reduced to %.> m&'( because AF>B of rigorously screened normal euthyroid volunteers have
serum TSH values between $.1 and %.> m&'(.
(b" TSH (ower <eference (imits
Cefore the immunometric assay era, TSH methods were too insensitive to detect values in the lower end of the
reference range (209). 7urrent methods however, are capable of measuring TSH at the lower end and now cite
lower limits between $.% and $.1 m&'( (202). !s the sensitivity of the methods has improved, there has been
an increased interest in defining the true lower limit of normal to better determine the presence of mild
(subclinical" hyperthyroidism. 7urrent studies suggest that TSH values in the $.. to $.1 m&'( range may
represent thyroid hormone e,cess and in elderly patients might be associated with an increased ris: of atrial
fibrillation, and cardiovascular mortality (36,37). t is therefore important to carefully e,clude subOects with a
goiter and any illness or stress in the normal cohort selected for reference range study.
- 34 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
,. -linical .se of Serum TSH /easurements
(a" Screening for Thyroid @ysfunction in !mbulatory ?atients
Most professional societies recommend that TSH be used for case finding or screening for thyroid dysfunction
in ambulatory patients, provided that the TSH assay used has a functional sensitivity at or below $.$% m&'(
(4,10,210). The TSH assay sensitivity stipulation is critical for the reliable detection of subnormal values, since
less sensitive TSH assays are prone to produce false negative (normal range" results on specimens with
subnormal TSH concentrations (202). The log'linear relationship between TSH and FT1 dictates that serum
TSH is the preferred test, since only TSH can detect mild (subclinical" degrees of thyroid hormone e,cess or
deficiency (Figure ." 5Section-% !.8. Mild (subclinical" thyroid dysfunction, characteri#ed by an abnormal TSH
associated with a normal range FT1 have reported prevalences in population surveys of K.$B and %B, for
subclinical hypo- and hyperthyroidism, respectively (10,18,25,211). @espite the clinical sensitivity of TSH, a
TSH-centered strategy has inherently two primary limitations. First, it assumes that hypothalamic-pituitary
function is intact and normal. Second, it assumes that the patients thyroid status is stable, i.e. the patient has had
no recent therapy for hypo-or hyperthyroidism 5Section-% !. and Figure %8 (19). f either of these criteria is not
met, serum TSH results can be diagnostically misleading (Table .".
)hen investigating the cause of an abnormal serum TSH in the presence of normal FT1 and FT6, it is important
to recogni#e that TSH is a labile hormone and subOect to nonthyroidal pituitary influences (glucocorticoids,
somatostatin, dopamine etc" that can disrupt the TSH'FT1 relationship (69,70,71,212). t is important to confirm
any TSH abnormality in a fresh specimen drawn after K6 wee:s before assigning a diagnosis of mild
(subclinical" thyroid dysfunction as the cause of an isolated TSH abnormality. !fter confirming a high TSH
abnormality, a T?3!b measurement is a useful test for establishing the presence of thyroid autoimmunity as the
cause of mild (subclinical" hypothyroidism. The higher the T?3!b concentration, the more rapid the
development of thyroid failure. !fter confirming a low TSH abnormality it can be difficult to uneGuivocally
establish a diagnosis of mild (subclinical" hyperthyroidism, especially if the patient is elderly and not receiving
(-T1 therapy (34). f a multinodular goiter is present, thyroid autonomy is the li:ely cause of mild (subclinical"
hyperthyroidism (213).
There is no consensus regarding the optimal age to begin the screening process. The !merican Thyroid
!ssociation guidelines recommend screening at age 6> and every > years thereafter (10). @ecision analysis
appears to support the cost-effectiveness of this strategy, especially for women (215). The strategy for using
TSH to screen for mild (subclinical" hypo- and hyperthyroidism will remain in dispute until there is more
agreement on the clinical conseGuences and outcome of having a chronically abnormal TSH. !lso there needs
to be agreement as to the level of the TSH abnormality that should indicate the need for treatment (216,217).
There is mounting evidence to suggest that patients with a persistent TSH abnormality may be e,posed to
greater ris: if left untreated. Specifically, a recent study reported a higher cardiovascular mortality rate when
patients had a chronically low serum TSH (37). Further, there are an increasing number of reports that indicate
that mild hypothyroidism in early pregnancy increases fetal wastage and impairs the Q of the offspring (63-
65)". Such studies support the efficacy of early thyroid function screening, especially in women during their
childbearing years.
(b" Nlderly ?atients
Most studies support screening for thyroid dysfunction in the elderly (10,35,214). The prevalence of both a low
and high TSH (associated with normal FT1" is increased in the elderly compared with younger patients.
Hashimotos+ thyroiditis, associated with a high TSH and detectable T?3!b, is encountered with increasing
prevalence as we get older. (35). The incidence of a low TSH is also increased in the elderly (35). ! low TSH
may be transient but is a persistent finding in appro,imately % B of elderly individuals, with no other apparent
evidence of thyroid dysfunction (36,214). This could be due to a change in the FT1'TSH set-point, a change in
TSH bioactivity or mild thyroid hormone e,cess (218). ! recent study by ?arle et al showed a higher
- 35 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
cardiovascular mortality rate in such patients (37). This suggests that the cause of a persistently low TSH level
should be actively investigated (37). Multinodular goiter should be ruled out as the cause especially in areas of
iodide deficiency (213). Medication history should be thoroughly reviewed (including over-the-counter
preparations, some of which contain T6". f a goiter is absent and the medication history negative, a serum TSH
should be rechec:ed together with T?3!b measurements after 1 to M wee:s. f the TSH is still low and T?3!b
is positive, the possibility of autoimmune thyroid dysfunction should be considered. Treatment of low TSH
should be made on a case-by-case basis.
(c" (-T1 <eplacement Therapy
t is now well documented that hypothyroid patients have serum FT1 values in the upper third of the reference
interval when the (-T1 replacement dose is titered to bring the serum TSH into the therapeutic target range ($.>-
%.$ m&'(" (219,220).
(evothyro,ine ((-T1" and not dessiccated thyroid, is the preferred long-term replacement medication for
hypothyroidism.
! euthyroid state is usually achieved in adults with a (-T1 dose averaging ..M Eg':g body weight'day. 7hildren
reGuire higher doses (up to 1.$Eg':g bw'day" and older individuals reGuire lower doses (..$ Eg':g bw'day"
(221,222). The initial dose and the optimal time needed to establish the full replacement dose should be
individuali#ed relative to age, weight and cardiac status. The reGuirements for an increase in thyro,ine during
pregnancy 5Section-% !68 and in post-menopausal women Oust starting hormone replacement therapy (223) may
also be increased.
! serum TSH result between $.> and %.$ m&'( is generally considered the therapeutic target for a standard (-
T1 replacement dose for primary hypothyroidism.
! serum FT1 concentration in the upper third of the reference interval is the therapeutic target for (-T1
replacement therapy when patients have central hypothyroidism due to pituitary and'or hypothalamic
dysfunction.
! typical schedule for gradually titrating to a full replacement dose involves giving (-T1 in %> Eg increments
each M /= wee:s until the full replacement dose is achieved (serum TSH $.>-%.$ m&'(". !s shown in figure %,
TSH is slow to re-eGuilibrate to a new thyro,ine level. ?atients with chronic, severe hypothyroidism may
develop pituitary thyrotroph hyperplasia which can mimic a pituitary adenoma, but which resolves after several
months of (-T1 replacement therapy (224). ?atients ta:ing <ifampin and anticonvulsants that influence the
metabolism of (-T1 may also need an increase in their dose of (-T1 to maintain the TSH within the therapeutic
target range.
Coth free T1 and TSH should be used for monitoring hypothyroid patients suspected of intermittent or non-
compliance with their (-T1 therapy. The parado,ical association of a high FT1 4 high TSH is often an
indication that compliance may be an issue. Specifically, acute ingestion of missed (-T1 doses before a clinic
visit will raise the FT1 but fail to normali#e the serum TSH because of the 9lag effect; (Figure %". n essence,
the serum TSH is analogous to the hemoglobin !.c as a long-term free T1 sensorP !t least M wee:s is needed
before retesting TSH following a change in the dose of (-T1 or brand of thyroid medication. !nnual TSH
testing of patients receiving a stable dose of (-T1 is recommended. The optimal time for TSH testing is not
influenced by the time of day the (-T1 dose is ingested (133). However, the daily dose should be withheld when
FT1 is used as the therapeutic endpoint, since serum FT1 is significantly increased (K.6B" above baseline for F
hours after ingesting the last dose (225).
deally (-T1 should be ta:en before eatingR at the same time of day and at least 1 hours apart from any other
medications or vitamins. Many medications can influence T1 absorption'metabolism (especially
7holestyramine, Ferrous Sulfate, Soy ?rotein, Sucralfate, antacids containing !luminum Hydro,ide,
anticonvulsants or <ifampin" (4,226).
- 36 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
(d" (-T1 Suppression Therapy
(-T1 administration designed to suppress serum TSH levels to subnormal values is typically reserved for
patients with well-differentiated thyroid carcinoma for which thyrotropin is considered a trophic factor (227).
The efficacy of (-T1 suppression therapy has been determined from uncontrolled retrospective studies that have
yielded conflicting results (228,229).
t is important to individuali#e the degree of TSH suppression by weighing patient factors such as age, clinical
status including cardiac factors and @T7 recurrence ris: against the potentially deleterious effects of iatrogenic
mild (subclinical" hyperthyroidism on the heart and bone (36). Many physicians use a serum TSH target of
$.$>-$.. m&'( for low-ris: patients and a TSH of -$.$. m&'( for high-ris: patients. Some physicians reduce
the (-T1 dose to give low-normal TSH values when patients have undetectable serum thyroglobulin (Tg" levels
and no recurrences >-.$ years after thyroidectomy. Suppression therapy for non-endemic goiters is generally
considered ineffective (230). Furthermore, patients with nodular goiters often already have suppressed TSH
concentrations as a result of thyroid gland autonomy (213).
Guideline !+. *evot0yro1ine 2*3T,4 )eplacement T0erapy for rimary Hypot0yroidism
(-T1, not desiccated thyroid, is the preferred medication for long-term replacement therapy for
hypothyroidism.
! euthyroid state is usually achieved with an average (-T1 dose of ..M Eg':g body weight'day. The initial
dose and time to achieve full replacement should be individuali#ed relative to age, weight and cardiac
status. !n initial (-T1 dose is normally >$-.$$ Eg daily. Serum TSH measurement after si, wee:s will
indicate the need for dose adOustment by %>->$ Eg increments.
7hildren reGuire higher doses of (-T1, up to 1.$Eg':g bw'day, due to rapid metabolism. Serum TSH and
FT1 values should be assessed using age-specific and method-specific reference ranges (Table 6".
! serum TSH level between $.> and %.$ m&'( is generally considered the optimal therapeutic target for
the (-T1 replacement dose for primary hypothyroidism.
TSH is slow to re-eGuilibrate to a new thyro,ine status (*uideline %". Si, to = wee:s is needed before
retesting TSH after changing the (-T1 dose or brand of thyroid medication.
ntermittent or non-compliance with levothyro,ine ((-T1" replacement therapy will result in discordant
serum TSH and FT1 values (high TSH'high FT1" because of a persistently unstable thyroid state (*uideline
%". Coth TSH and FT1 should be used for monitoring such patients.
Thyro,ine reGuirements decline with age. 3lder individuals may reGuire less than ..$ Eg':g bw'day and
may need to be titrated slowly. Some physicians prefer to gradually titrate such patients. !n initial dose of
%> Eg is recommended for patients with evidence of ischemic heart disease followed by dose increments of
%> Eg every 6-1 wee:s until the full replacement dose is achieved. Some believe that a higher target TSH
($.>-6.$ m&'(" value may be appropriate for the elderly patient.
n severe hypothyroidism an initial (-T1 loading dose is the most rapid means for restoring a therapeutic
FT1 level because the e,cess of unoccupied binding sites may blunt the FT1 response to treatment.
Thyro,ine reGuirements increase during pregnancy. Thyroid status should be chec:ed with TSH 4 FT1
during each trimester of pregnancy. The (-T1 dose should be increased (usually by >$ Eg'day" to maintain
a serum TSH between $.> and %.$ m&'( and a serum FT1 in the upper third of the normal reference
interval.
?ost-menopausal women starting hormone replacement therapy may need an increase in their (-T1 dose to
:eep the serum TSH within the therapeutic target.
TSH testing of patients receiving a stable (-T1 dose is recommended on an annual basis. The ideal time for
TSH testing is not influenced by the time of day the (-T1 dose is ingested.
deally (-T1 should be ta:en before eating, at the same time of day, and at least 1 hours apart from any
other medications or vitamins. Cedtime dosing should be % hours after the last meal.
?atients beginning chronic therapy with cholestyramine, ferrous sulfate, calcium carbonate, soy protein,
sucralfate and antacids containing aluminum hydro,ide that influence (-T1 absorption may reGuire a larger
(-T1 dose to maintain TSH within the therapeutic target range.
- 37 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
?atients ta:ing <ifampin and anticonvulsants that influence the metabolism of (-T1 may also need an
increased (-T1 dose to maintain the TSH within the therapeutic target range.
Guideline !,. *evot0yro1ine 2*3T,4 Suppression T0erapy
Serum TSH is considered a growth factor for differentiated thyroid cancer (@T7". The typical (-T1 dose
used to suppress serum TSH in @T7 patients is %..Eg':g body weight'day.
The target serum TSH level for (-T1 suppression therapy for @T7 should be individuali#ed relative to the
patient+s age and clinical status including cardiac factors and @T7 recurrence ris:.
Many physicians use a serum TSH target value of $.$>-$.. m&'( for low-ris: patients and a TSH of -$.$.
m&'( for high-ris: patients.
Some physicians use a low-normal range therapeutic target for TSH when patients have undetectable serum
Tg levels and have had no recurrence >-.$ years after thyroidectomy.
f iodide inta:e is sufficient, (-T1 suppression therapy is rarely an effective treatment strategy for reducing
the si#e of goiters.
3ver time, multi-nodular goiters typically develop autonomy that is characteri#ed by a subnormal serum
TSH level. Serum TSH should be chec:ed before initiating (-T1 suppression therapy in such patients.
.i( 4& T%e seru TSH5.T4 relations%i# t"#ical o' di''erent clinical conditions
- 38 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
(e" Serum TSH Measurement in Hospitali#ed ?atients with 0T
!lthough most hospitali#ed patients with 0T have normal serum TSH concentrations, transient TSH
abnormalities in the $.$% / %$ m&'( range are commonly encountered in the absence of thyroid dysfunction
(20,87,92,93). t has been suggested that the use of a wider reference range ($.$% /.$ m&'(" would improve
the positive predictive value of TSH measurements for evaluating the sic: hospitali#ed patient (20,92,93,231).
TSH should be used in conOunction with a FT1 estimate (or TT1" test for evaluating hospitali#ed patients with
clinical symptoms or patients with a history of thyroid dysfunction (*uidelines M and %>".
Sometimes the cause of the TSH abnormality in a hospitali#ed patient is obvious, such as in the case of
dopamine or glucocorticoid-therapy (87,92). n other cases the TSH abnormality is transient and appears to be
caused by 0T per se, and resolves with recovery from the illness. t is common to see a transient minor
suppression of TSH into the $.$%-$.% m&'( range during the acute phase of an illness, followed by a rebound
to mildly elevated values during recovery (103). t is important to use a TSH assay with a functional sensitivity
2 $.$% m&'( in the hospital setting in order to be able to reliably determine the degree of TSH suppression.
Specifically, the e,tent of TSH suppression can be used to discriminate sic: hyperthyroid patients with
profoundly low serum TSH values (- $.$% m&'(" from patients with a mild transient suppression of 0T (20).
Guideline !5. TSH /easurement in Hospitali6ed atients
TSH 4 T1 (FT1 or TT1" is the most useful test combination to detect thyroid dysfunction in a sic:
hospitali#ed patient.
t is more appropriate to use a widened TSH reference interval ($.$> to .$.$ m&'(" in the hospitali#ed
setting. Serum TSH levels may become subnormal transiently in the acute phase and become elevated in
the recovery phase of an illness.
! serum TSH value between $.$> and .$.$ m&'( is usually consistent with a euthyroid state, or only a
minor thyroid abnormality that can be evaluated by retesting after the illness subsides. (This only applies to
patients not receiving medications such as dopamine that directly inhibit pituitary TSH secretion."
! low-normal TSH level in the presence of a low TT1 and low TT6 may reflect central hypothyroidism as a
result of a prolonged illnessR whether or not this condition reGuires immediate treatment remains uncertain
and is currently controversial.
)hen thyroid dysfunction is suspected, a thyroid pero,idase antibody (T?3!b" test may be useful to
differentiate autoimmune thyroid disease from 0T.
@iagnosing hyperthyroidism in 0T patients can be a challenge because current FT1 methods can give both
inappropriately low and high values in euthyroid 0T patients (101,232). Serum TT1 and TT6 measurements
may be useful for confirming a diagnosis of hyperthyroidism if assessed relative to the severity of the illness
(*uideline M". ! suppressed serum TSH below $.$%m&'( is less specific for hyperthyroidism in hospitali#ed
individuals, compared with ambulatory patients. 3ne study found that .1B of hospitali#ed patients with TSH -
$.$$> m&'( were euthyroid. However, such patients had a detectable T<H-stimulated TSH response whereas
patients who were truly hyperthyroid with 0T did not (20).
Mild (subclinical" hypothyroidism cannot be easily diagnosed during a hospitali#ation, because of the freGuency
of high TSH abnormalities associated with 0T. ?rovided that the thyroid hormone (FT1 or TT1" concentration
is within normal limits, any minor abnormality in serum TSH ($.$%-%$.$ m&'(" arising from a mild
(subclinical" thyroid condition is unli:ely to affect the outcome of the hospitali#ation, and can be deferred for
evaluation %-6 months after discharge. n contrast, sic: hypothyroid patients typically e,hibit the combination
of low FT1 and elevated TSH (A%$ m&'(" (92).
(f" 7entral Hypothyroidism
The log'linear relationship between TSH and FT1 dictates that patients with primary hypothyroidism and a
subnormal FT1 should have a serum TSH value A .$m&'( (Figure ." 5Section-% !.8. )hen the degree of TSH
- 39 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
elevation in response to a low thyroid hormone level appears inappropriately low, pituitary insufficiency should
be e,cluded. ! diagnosis of central hypothyroidism will usually be missed using a 9TSH first; strategy (19).
Guideline !7. *evot0yro1ine 2*3T,4 )eplacement T0erapy for -entral Hypot0yroidism
! serum FT1 level in the upper third of the reference interval is the therapeutic target for the (-T1
replacement dose used to treat central hypothyroidism due to pituitary or hypothalamic dysfunction.
)hen using FT1 as the therapeutic endpoint for central hypothyroidism, the daily dose of (-T1 should be
withheld on the day of the FT1 measurement. (Serum FT1 is increased (K.6 B" above baseline for F hours
after ingesting (-T1".
7entral hypothyroid conditions are characteri#ed by parado,ically normal or slightly elevated serum TSH in the
maOority of cases (29). n one study of central hypothyroid patients, 6>B had subnormal TSH values but 1.B
and %>B had inappropriately normal and elevated TSH values, respectively (233). t is now well documented
that the parado,ically elevated serum TSH levels seen in central hypothyroid conditions is caused by the
measurement of biologically inert TSH isoforms that are secreted when the pituitary is damaged or when
hypothalamic T<H stimulation is deficient (197). The inappropriate TSH values arise because the monoclonal
antibodies used in current TSH assays cannot distinguish between TSH isoforms of different biological activity,
since TSH biological activity is determined not by the protein bac:bone but by the degree of glycosylation,
specifically the sialylation of the TSH molecule. t appears that normal T<H secretion is essential for producing
normal TSH sialylation and association of the TSH subunits to form mature, biologically active TSH molecules
(29,197,234). The biological activity of TSH in central hypothyroid conditions appears to be inversely related to
the degree of TSH sialylation as well as the FT1 level in the circulation (29). T<H-stimulation testing may be
useful for specifically diagnosing central hypothyroidism (235). Typical TSH-responses in such conditions are
blunted (-%-fold rise' 21.$ m&'( increment" and may be delayed (197,204,235,236). n addition, the T6-
responses to T<H-stimulated TSH are blunted and correlate with TSH biological activity (197,237,238).
Guideline !8. -linical .tility of TSH Assays 2#unctional Sensitivity 2 $.$! mI.9*4
Serum TSH measurement is the most diagnostically sensitive test for detecting mild (subclinical", as well as
overt, primary hypo- or hyperthyroidism in ambulatory patients.
The maOority (AF>B" of healthy euthyroid subOects have a serum TSH concentration below %.> m&'(.
!mbulatory patients with a serum TSH above %.> m&'( when confirmed by a repeat TSH measurement
made after 6-1 wee:s, may be in the early stages of thyroid failure, especially if T?3!b is detected.
! serum TSH measurement is the therapeutic endpoint for titrating the (-T1 replacement dose for primary
hypothyroidism (see *uideline %6" and for monitoring (-T1 suppression therapy for differentiated thyroid
carcinoma (see *uideline %1".
Serum TSH measurements are more reliable than FT1 in hospitali#ed patients with non-thyroidal illness not
receiving dopamine. Serum TSH should be used in conOunction with T1 (TT1 or FT1" testing for
hospitali#ed patients (*uidelines M and %M".
TSH cannot be used to diagnose central hypothyroidism because current TSH assays measure biologically
inactive TSH isoforms.
7entral hypothyroidism is characteri#ed by an inappropriately normal or slightly elevated serum TSH level
and a blunted (-%-fold rise' 21.$ m&'( increment" T<H response.
)hen the serum FT1 is low and yet the serum TSH is only minimally elevated (-.$ m&'(", a diagnosis of
central hypothyroidism should be considered.
Serum TSH measurements are an important pre-natal and first trimester screening test to detect mild
(subclinical" hypothyroidism in the mother (see *uideline 1".
! low TSH in the setting of a multinodular goiter suggests the presence of mild (subclinical"
hyperthyroidism due to thyroid autonomy.
! serum TSH measurement is reGuired for confirming that an elevated thyroid hormone level is due to
hyperthyroidism and not a thyroid hormone binding protein abnormality (such as F@H".
! serum TSH measurement is the primary test for detecting amiodarone / induced thyroid dysfunction (see
*uideline >".
- 40 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
(g" nappropriate TSH Secretion Syndromes
!s shown in Table ., binding protein abnormalities or assay technical problems are the most common causes
for a discordant FT1'TSH relationship. The apparent parado,ical dissociation between high levels of thyroid
hormone and a non-suppressed serum TSH has led to the widespread use of the term 9inappropriate TSH
secretion syndromes; to describe these conditions. Specimens that display a discordant TSH'FT1 relationship
are increasingly being identified now that sensitive TSH assays that can reliably detect subnormal TSH
concentrations are available and in widespread use 5Section-6 7%8. !s shown in Table ., it is critical to first
e,clude li:ely causes of a discordant TSH'FT1 ratio, i.e technical interference and'or binding protein
abnormality. This confirmatory testing should be performed on a fresh specimen by measuring TSH together
with free and total thyroid hormone with a different manufacturer+s method. 3nly after the more common
causes of discordance have been eliminated should rare conditions such as a TSH-secreting pituitary tumor or
thyroid hormone resistance be considered.
)hen the abnormal biochemical profile has been confirmed, the possibility that a TSH-secreting pituitary tumor
is the cause of the parado,ical TSH should first be eliminated before assigning the diagnosis of thyroid
hormone resistance (TH<". t should be noted that both conditions can coe,ist (247). TSH-secreting pituitary
tumors have similar biochemical profiles to TH< but can be distinguished from TH< by TSH-alpha subunit
testing and radiographic imaging. !dditionally, T<H-stimulation testing may occasionally be useful in
developing the differential diagnosis. Specifically, a blunted T<H-stimulation test and T6-suppression test is
characteristic of most TSH-secreting pituitary tumors whereas a normal response is seen in most cases of TH<
(245).
!i" TSH-Secreting #ituitary Tumors
?ituitary tumors that hypersecrete TSH are rare, representing -.B of cases of inappropriate TSH secretion
(27,28). These tumors often present as a macroadenoma with symptoms of hyperthyroidism associated with a
non-suppressed serum TSH and M< evidence of a pituitary mass (28).
!fter e,cluding a technical reason for the parado,ically elevated TSH level (i.e. H!M!", the diagnosis of TSH-
secreting pituitary tumor is usually made on the basis ofH
! lac: of TSH response to T<H stimulation
!n elevated serum TSH alpha subunit
! high alpha subunit'TSH ratio
The demonstration of a pituitary mass on M<
Guideline !8. #or /anufacturers of TSH Tests
Manufacturers that mar:et TSH tests with differing sensitivities are urged to discontinue mar:eting the less
sensitive product.
There is no Oustification for the pricing of TSH assays to be based on sensitivityP
There is no scientific Oustification for refle,ing from a less sensitive to a more sensitive TSH test.
Manufacturers should help laboratories independently establish functional sensitivity by providing
appropriately low TSH human serum pools when reGuested.
Manufacturers should indicate the use of calibration factors, especially if calibration factors are country-
dependent.
Manufacturers should Guote the recovery of the TSH <eference ?reparation at the claimed functional
sensitivity.
Sit ?ac:age nserts shouldH
- @ocument the realistic functional sensitivity of the method using the *uideline %$ protocol
- 7ite the functional sensitivity that can be achieved across a range of clinical laboratories
- @isplay the typical between-run precision profile e,pected for a clinical laboratory
- <ecommend the use of functional sensitivity not analytical sensitivity to determine the
- 41 -
NACB: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease
Laurence M. Demers, h.D., !.A.C.B.and Caro"e A. Spencer h.D., !.A.C.B.
lowest reporting limit. (!nalytical sensitivity prompts laboratories to adopt an unrealistic
detection limit."
!ii" Thyroid Hormone $esistance
Thyroid hormone resistance (TH<" is usually caused by a mutation of the thyroid hormone (T<", T<-beta
receptor gene that occurs in .H >$,$$$ live births (239-242). !lthough the clinical presentation can be variable,
patients have a similar biochemical profile. Specifically, serum FT1 and FT6 are typically elevated (from a
minimal degree to a %-6-fold elevation above the upper normal limit" and associated with a normal or slightly
elevated serum TSH that responds to T<H stimulation (242,243). However, it should be recogni#ed that TSH
secretion is not inappropriate given the fact that the tissue response to thyroid hormone is reduced, reGuiring
higher thyroid hormone levels to maintain a normal metabolic state. TH< patients typically have a goiter as a
result of chronic hypersecretion of a hybrid TSH isoform that has increased increased biologic potency
(199,244). The clinical manifestation of thyroid hormone e,cess covers a wide spectrum. Some patients appear
to have a normal metabolism with a near-normal serum TSH and whose receptor defect appears to be
compensated for by high levels of thyroid hormone (*enerali#ed TH<". 3ther patients appear to be
hypermetabolic and to have a defect that selectively affects the pituitary (?ituitary TH<".
The distinctive features of TH< are the presence of a non-suppressed TSH, together with an appropriate
response to T<H despite elevated thyroid hormone levels (242,245). !lthough rare, it is important to consider
the diagnosis of TH< when encountering a patient with elevated thyroid hormone levels associated with a
parado,ically normal or elevated TSH (242,246). Such patients have often been misdiagnosed as having
hyperthyroidism and have been subOected to inappropriate thyroid surgery or radioiodide gland ablation (242).
- 42 -