Scand. J. clin. Lab. Invest. 39, 7-13, 1979. Unconjugated thyroxine and triiodo t hyronine in urine: influence of age, sex, drugs and thyroid function E. IVERSEN & K. 0. PEDERSEN Department of Internal Medicine and Department of Clinical Chemistry, Skive Sygehus, DK 7800 Skive, Denmark Iversen, E. & Pedersen, K.O. Unconjugated thyroxine and triiodothyronine in urine: influence of age, sex, drugs and thyroid function. Scand. J. din. Lab. Invest. 39, 7-13, 1979. The 24 h urinary excretion (dU) of T4 and T3, determined by radioimmuno- assays using extraction and separation on Sephadex columns, is higher in men than in women (mean dU-T4 and dU-T3f2 SD: 2.6f0.9 vs. 2.2fl.O and 2.5 k 1.0 vs. 2.0 f0.8 nmol, respectively). The excretion of T3 decreases with age, but expressed relatively to dU-creatinine there is no effect of age and sex. Children have increased T4 and T3 to creatinine excretion ratios and neonates excrete predominantly T4. Oestrogens, acute salicylate loading and diurnal variation yielded results, which were at variance with the hypothesis that dU-T4 and dU-T3 mirror corresponding free hormone concentrations in serum, and there was only a weak positive correlation to total serum hormones in thyro- toxicosis as well. General clinical use of urine T4 and T3 is obviated by poor diagnostic discrimination and inherent analytical and interpretative disadvant- ages. Key-words: diurnal variation; myxoedema; radioimmunoassay ; renal function ! salicylate; Sephadex; thyrotoxicosis Eigil Iuersen, M. D., Second University Clinic of Internal Medicine, Kommune- hospitalet, 8000 Aarhus C , Denmark Since the first reports on urinary excretion of thyroxine (T4) and triiodothyronine (T3) as promising indices of thyroid function [6, 8, 91, there has been disagreement between published reference values [4, 6, 8, 9, 13, 14, 21, 221 and conflicting views upon the clinical usefulness of these quantities [6, 8, 9, 10, 14, 16, 17, 21, 221. However, the prevailing view is still that excreted hormones represent integrated para- meters for the corresponding free hormone concentrations in blood during the period of excretion [13, 141. 0036-551 3/79/0200-0007 $02.00 0 1979 Medisinsk Fysiologisk Forenings Forlag It is the purpose of this paper to examine the T4 and T3 excretion in more detail as function of age and drugs affecting thyroxine binding globulin (TBG) concentration and binding capacity, and to evaluate the usefulness of these two quantities in the diagnosis of hyper- and hypofunction of the thyroid. MATERIAL AND METHODS Reference inaferial 24 h urine collections were obtained from the following: (1) sixty-seven healthy euthyroid 7 Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 04/30/13 For personal use only.
Transcript
Scand. J. clin. Lab. Invest. 39, 7-13, 1979.
Unconjugated thyroxine and t riiodo t hyronine in urine: influence of age, sex, drugs and thyroid function E. IVERSEN & K . 0. P E D E R S E N Department of Internal Medicine and Department of Clinical Chemistry, Skive Sygehus, DK 7800 Skive, Denmark
Iversen, E. & Pedersen, K.O. Unconjugated thyroxine and triiodothyronine in urine: influence of age, sex, drugs and thyroid function. Scand. J . d i n . Lab. Invest. 39, 7-13, 1979.
The 24 h urinary excretion (dU) of T4 and T3, determined by radioimmuno- assays using extraction and separation on Sephadex columns, is higher in men than in women (mean dU-T4 and dU-T3f2 SD: 2.6f0.9 vs. 2 .2f l .O and 2.5 k 1.0 vs. 2.0 f0.8 nmol, respectively). The excretion of T3 decreases with age, but expressed relatively to dU-creatinine there is no effect of age and sex. Children have increased T4 and T3 to creatinine excretion ratios and neonates excrete predominantly T4. Oestrogens, acute salicylate loading and diurnal variation yielded results, which were at variance with the hypothesis that dU-T4 and dU-T3 mirror corresponding free hormone concentrations in serum, and there was only a weak positive correlation to total serum hormones in thyro- toxicosis as well. General clinical use of urine T4 and T3 is obviated by poor diagnostic discrimination and inherent analytical and interpretative disadvant- ages.
Eigil Iuersen, M. D., Second University Clinic of Internal Medicine, Kommune- hospitalet, 8000 Aarhus C, Denmark
Since the first reports on urinary excretion of thyroxine (T4) and triiodothyronine (T3) as promising indices of thyroid function [6, 8, 91, there has been disagreement between published reference values [4, 6, 8, 9, 13, 14, 21, 221 and conflicting views upon the clinical usefulness of these quantities [6, 8, 9, 10, 14, 16, 17, 21, 221. However, the prevailing view is still that excreted hormones represent integrated para- meters for the corresponding free hormone concentrations in blood during the period of excretion [13, 141.
It is the purpose of this paper to examine the T4 and T3 excretion in more detail as function of age and drugs affecting thyroxine binding globulin (TBG) concentration and binding capacity, and to evaluate the usefulness of these two quantities in the diagnosis of hyper- and hypofunction of the thyroid.
MATERIAL A N D M E T H O D S
Reference inaferial
24 h urine collections were obtained from the following: (1) sixty-seven healthy euthyroid
7
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
8 E. Iversen & K. 0. Pedersen
hospital employees; (2) sixty-one adult medical inpatients, who had no clinical signs of thyroid disease or biochemical signs of renal disease, were afebrile and received no medication known to interfere with thyroid function; and (3) from fourteen healthy children, aged 2-10 years. A spot urine sample was obtained from eighteen newborn children on their fourth day of living after uncomplicated birth.
The age ranges and medians (in parenthesis) for groups 1 , 2 and 3 were: 20-57 (30), 20-69 (49) and 2-10 (7) years, respectively. The number of men and women (in parenthesis) in the five decades 20-70 years were: 17 (24), 16 (19), 4 (1 l), 10 (9) and 1 1 (6) respectively.
Experimental material
(1) 24 h urine collections were obtained from twelve women (mean age 27 years), who for more than 3 months had taken oestrogen con- taining contraceptive pills (none were low oestro- gen dose pills).
(2) On the day after a 24 h urine collection 3 g enteric coated acetyl salicylic acid was given orally to nine healthy persons, who then col- lected four consecutive 6 h urines.
(3) Urine was collected during four consecu- tive 6 h periods from fifteen healthy ambulatory persons and from eight patients confined to bed with minor orthopaedic diseases.
Thyroid material
24 h urine collections were obtained from nineteen and five patients with biochemical and clinical signs of thyrotoxicosis and myxo- edema, respectively. The age ranges and medians (in parenthesis) were 32-81 (61) years and 52-76 years (59), respectively. All the patients were seen by one of us (E.I.).
No persons in the three groups had protein- uria, as revealed by Albustix" strips. Urinary creatinine was determined in every urine sample as a check upon collection error and a well mixed aliquot was frozen for hormone deter- mination immediately after termination of the collecting period and measurement of volume.
Urinary T4 and T3 methods
Reagents. T4 and T3 stock standards, 300 nmol/l, were made from respective free acids
(Sigma) in solutions containing HSA (Kabi) 2 g/l and NaCl 155 mmol/l. Working standards covering the range from 0 to 7.3 nmol/l were made before each assay by appropriate dilution of the stock standards by a 155 mmol/l NaCl solution.
Labelled hormones ([1z51]T4 and ['251]T3) with specific activities of 39-64 and 46-77 Ci/mmol, respectively, were obtained monthly from the Radiochemical Centre, Amersham, England. Working isotope solutions were made before each assay by dilution 20,000 times in a 0.08 mol/l NaOH solution containing HSA 1 g/l. Specific T4 and T3 antibodies (kindly donated by V. Kruse, National Institute of Animal Science, Copenhagen, and A. G. Veje, Department of Clinical Chemistry, Holbzek, respectively) were diluted 4,000 and 17,000 times, respectively, in a 75 mmol/l barbital buffer, pH 8.5, before use.
The barbital buffer was made by dissolving 15.5 g diethyl barbituric acid, sodium salt (Merck) in demineralized water with ensuing adjustment of pH to 8.5 with 1 mol/l HCI, whereafter volume was made to 1 1 with de- mineralized water.
Procedure. Apart from reagent concentrations the following procedure was for practical reas- ons identical to the one used in this laboratory for determination of T3 in serum. Small re- usable Sephadex G-25 fine columns (Tetralute@, Ames Company) were emptied by pouring off NaOH solution above the surface of the columns, and placed in a special rack. 200 p1 standard, control urine or urine sample was diluted with 400 p1 working isotope dilution by a Hook & Tucker autodilutor and delivered directly on thesurface(with bottomcap removed) and allowed to drain into the column. Fifteen minutes later 4 ml barbital buffer was added to the column, which was allowed to drain completely. The effluents of these two steps were discarded and the rack with columns was then placed above a matched rack with counting tubes. Four hundred microlitres of the appropriate antibody working solution was added and antibody bound hormone was eluted after 2 h incubation at room temperature by addition of 4 ml barbital buffer. The com- bined effluents of the two last steps were counted for 5 min and results were calculated on a programmable calculator using a linear cor-
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
T4 and T3 in human urine 9
relation (r> 0.99) between standard con- centration and reciprocal of counts. All determinations were performed in duplicate. The columns were cleaned and stored as described elsewhere [21].
Other methods
Serum T4 and T3 were measured by the same Sephadex column method employing dilute human serum as binder for T4 [2] and the T3 antibody used above as binder for T3 [3]. The free thyroxine index in serum was obtained by multiplying the serum T4 con- centration by the value for a modified T3 Sephadex uptake test [15]. Urine creatinine was measured by an alkaline picrate method [5].
RESULTS
Methodological
The analytical performance of the urine T4 and T3 methods are given in Table I, which documents reasonable precision, sensitivity, recovery and negligible mutual interference. Addition of HSA and y-globulin (10% solution from the State Serum Institute, Copenhagen) up to 5 g/l urine and sodium salicylate up to 5 mmolil urine had no effect upon T4 or T3 concentration in the urine.
Biological
Reference values. Table I1 shows mean values and standard deviations for dU-T4 and dU-T3 in persons of reference, experimental and thyroid groups. Men excreted significantly
more T4 and T3 than women (P < 0.025 and <0.01, respectively, t test). However, when T4 and T3 excretions were expressed relatively to creatinine excretion, there was no significant sex difference. The ratio dU-T4/dU-T3 was equal in both sexes.
The effect of age is shown in Fig. 1 . There is a significant decrease of dU-T3 in men and women with age, but no age relation, when dU-T3 is expressed relatively to dU-creatinine. The excretion of T4 in women over 30 years of age decreased both absolutely and relatively, whereas the T4 excretion in men showed no consistent pattern.
Children excreted less T4 than adults (P< 0.025) but the same amount of T3. In contrast, newborn children had very low T3 concentration (below the detection limit in seven of eighteen samples), but T4 concentration at adult level. The T4 and T3 to creatinine excretion ratios in children increased with their decreasing creatin- ine excretion.
Effect of TBG concentration and binding. Women on oestrogen containing contraceptive pills had significantly higher T4 excretion (P < 0.001) whereas their T3 excretion showed no significant change in comparison with the reference ranges of the matched age groups. Their dU-T4/dU-T3 ratio was significantly higher (P < 0.02).
The acute oral salicylate loading experiment showed a significant decrease of T3 excretion during the ensuing three 6 h periods, whereafter excretion returned to control level, as shown in Table 111. The T4 excretion showed a similar but less dramatic and more variable decrease and recovery.
A control experiment showed no diurnal
TABLE I. Analytical performance of methods for urine T4 and T3
Urine T4 method Urine T3 method
Long term precision; 5.6% 5.2% SD on duplicate determinations** 0.32 nmol/l 0.27 nmol/l Detection limit? 0.33 nmol/l 0.21 nmolll Interference from T31: 0.8% Interference from T4 - Recovery of added T4 and T3 (n = 10) 97+ 16%
0.4% 99k 6%
* In fourteen aliquots of a frozen urine pool during 4 months (mean T4 and T3 were 8.6
** In 122 consecutive urine samples with T4 and T3 concentrations up to 13 nmol/l. t Defined as three times SD of the zero standard (n = 10). $ From ref. [IS].
and 3.8 mmol/l, respectively).
B
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
10 E. Iversen & K . 0. Pedersen
TABLE 11. Summary of mean values, standard deviations and ranges for urine iodothyronine excretion in reference material and patients during 24 h
U-T4/U- U-T3/U- U-T4 nmol U-T3 nmol Creat.x lo7 Creat.x l o 7 Ratio
n (mean? SD) (mean+ SD) (nieanf SD) (mean+ SD) U-T4/U-T3
Men 58 2.58k0.86 Women 69 2.21k0.96 Children 14 1.76kO.73 Neonates 18 2.26*
Women on oestrogens 12 3.73k1.23 Thyreotoxicosis 19 6.19k4.95
* Indicates concentration values. ** Calculated from the respective mean values. The values in parentheses are the absolute ranges. n is number of
persons or patients. ND, Not detectable ( 5 0 . 3 nmol for T4, 50 .2 nniol for T3).
variation of T3 excretion in ambulatory and supine persons (Table TV), whereas the T4 excretion was about 20% higher during the morning than during the night in both ambula- tory and supine persons.
Valiles iti thyroid disease. The nineteen patients with thyrotoxicosis had increased and the five patients with myxoedema had decreased mean
values for dU-T4 and dU-T3, but the ratio between mean T4 and mean T3 excretion was similar to the reference group (Table 11). None of the patients had isolated T3 thyro- toxicosis.
Among the thyrotoxic patients five (mean age 71 years) and four (mean age 73 years) had values within the reference ranges for dU-T4 and dU-T3 above 60 years, respectively. Three
hv 40 ' ' 60 ' 80 ' %o ' 40 ' ' 60 ' ' 80 I
Age (years)
FIG. 1. Effect of age upon 24 h excretion of T4 (dU-T4) and T3 (dU-T3) and upon the excretion iacios between T4 and T3 and creatinine. Males, ; females, 0. Vertical lines designate mean values& SEM.
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
T4 and T3 in human urine 11
TABLE 111. Effect of a single oral dose (3 g) of acetyl salicylic acid on urinary excretion of T4 and T3 in nine healthy subjects
Time period % T4 excreted % T3 excreted after dose (h) (mean+ SD) (mean+ SD)
0-6 106? 26 76+ 18** 6-1 2 56+ 36* 41 +21***
12-18 70i 33** 48+ 16*** 18-24 98+27 90+20
~ ~-
Significance of differences between control and experimental period: * P<O.OS; ** P<O.Ol; * * * P i 0.001. T4 and T3 excreted are expressed as percentage of mean 6 h excretion in the preceding 24 h. There was no significant change in creatinine excretion during the four periods or in comparison with the preceding 24 h.
patients (mean age 80 years) had normal values for both quantities. Among the patients with myxoedema three and two had values below the detection limit for T4 and T3, respectively, and both hormones were undetectable in one patient. Two patients had values within the reference range for dU-T3 and two other pati- ents within the reference range for dU-T4.
Comparative analysis of parameters. Com- parison with linear regression analysis in the thyrotoxic group showed an insignificant positive correlation between S-T4 ( r = 0.10) and S-free T4 index (r = 0.1 1) on the one hand
TABLE IV. Diurnal variation of T4 and T3 excretion in urine
% T4 excreted? SD % T3 excreted+ SD Time period ambulatory (n = 15) Supine (n = 8) ambulatory (n = 15) Supine (n = 8)
Significance of difference between consecutive pairs (paired r-test): * P<O.O05; ** P<O.OS. T4 and T3 excreted has been expressed as percentage of mean excretion in 6 h. There was no significant difference between the creatinine excretions in the four period.
330 150
i
170 70
130 r - - 0
-
.I * * 1 . I _ _ _ - - - - I
35/
0
t
( "
0
1 0 I I I *
0
Diagnostic index score
FIG. 2. Correlation between diagnostic index score [ l l ] and S-T4 (0) and S-free T4 index (0) in the left half and 24 h excretion of T4 (dU-T4, 0 ) and T3 (dU-T3, 0) in the right half of the figure. The horizontal lines designate the upper boundaries of respective reference ranges and the vertical stippled line the boundary between equivocal and definite thyrotoxic range [I I].
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
12 E. Iversen & K. 0. Pedersen
and dU-T4 on the other. S-T3 and dU-T3 were significantly positively correlated (r = 0.48, P<O.O5) and dU-T3 and dU-T4 were highly significantly correlated (r = 0.97, P <0.001). A comparison with clinical impression as quantitated by a diagnostic index score based on the presence and absence of symptoms and signs of thyrotoxicosis [I I ] showed an insignifi- cant positive correlation to S-T4 ( r = 0.10), S-T3 (r = 0.24), S-free T4 index (r = 0.14) and dU-T3 (r = 0.40), but a significant positive correlation to dU-T4 (r = 0.46, P < 0.05). The diagnostic score was significantly negatively correlated to age (r = -0.65, P<O.OOI) , whereas there was no significant correlation between age and serum or urine values. Fig. 2 shows how clinical severity of thyrotoxicosis correlates with serum and urine values in relation to upper normal boundaries for all parameters.
DISCUSSION
T4 and T3 are excreted in urine as free, protein bound and conjugated hormones [6, 7, 13, 141, the latter mainly as glucuronides and sulphates. The four to six fold variation in mean values for dU-T4 and dU-T3 reported in the literature reflects variable hydrolysis of conjugates during the extraction step [6] and adverse effect of even the slightest degree of proteinuria on the separation step in some methods [7, 221.
In agreement with two recent works using the same principle [14, 211 the present method measures only or mainly the free and the protein bound moities and is unaffected by proteinuria per se. These three works agree reasonably upon reference values for dU-T4; the three-fold variation of reference values for dU-T3, how- ever, can hardly be due to demographic differ- ences, but must include methodological bias due to non-specificity of T3 antibodies [lo] or radiochemical impurities in the labelled hormones. The latter source of error was at a minimum with the present methods, which employed hormones of low specific activity, but we have not established to what extent T3 or T4 metabolites crossreact with T3 antibody.
As regards reference values, the significant decrease of dU-T3 with age has been reported earlier [20], and it coincides with the decrease in glomerular filtration rate and muscle mass
with age, so that the excretion ratio between T3 and creatinine is constant and unaffected by both age and sex.
Nothing is known about the relative import- ance of filtration, reabsorption, secretion or deiodination (of T4 to T3 [l]) as sources of T4 and T3 in urine. The findings in this work are at variance with the hypothesis, that urinary T4 and T3 mirrors the corresponding free hormone concentrations in plasma since : there was only a weak positive correlation between serum and urine values in patients with thyrotoxicosis; an oestrogen-induced rise of TBG concentration increased only the T4 excretion, whereas the concentration of free hormones in serum is unaffected by oestrogens; an acute increase of the free hormone fractions in serum produced by an oral salicylate dose [ 191 lead to an unexpected decreased excretion of both hormones; and finally the excretion of T3 is constant and the excretion of T4 is lowest during night, when both free hormone con- centrations in serum are highest [23]. The diurnal variation of T4 excretion was identical in ambulatory and supine persons, thereby making slight proteinuria an unlikely explana- tion [14, 221, and the more variable response of T4 excretion upon salicylate loading can thus partly be explained by the superimposed diurnal variation.
The independence of the T3 to creatinine excretion ratio from age, sex and diurnal variation should in theory make determination of this quantity in a random urine sample mailed to the laboratory a convenient test of thyroid function. However, the Iack of correla- tion to free hormones in serum, the insufficient diagnostic discrimination (admittedly this con- clusion involves a circular argument), the wide reference ranges, inherent specificity problems, unpredictable response to drugs, and finally, the required absence of proteinuria and know- ledge of renal function [12] all point to measure- ment of thyroid hormones in serum as the more attractive alternative. This conclusion is hardly surprising, when it is considered that less than 5 % of the total daily production of T4 and T3 is excreted unchanged in the urine, and the apparently better fit between clinical impression and urine hormones in thyrotoxicosis is inter- preted as fortuitious since the paucity of symptoms and signs in our predominantly old patients coincided with the decreasing hormone
Scan
d J
Clin
Lab
Inv
est D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y N
yu M
edic
al C
ente
r on
04/
30/1
3Fo
r pe
rson
al u
se o
nly.
T4 and T3 in human urine 13
excretion with age. The effect of other drugs on renal excretion of T4 and T3 and the character- istic excretion pattern in newborn children deserve more study.
A C K N O W L E D G M E N T
This study was supported by Overlzge Skat Baastrups Jubil~eumslegat.
R E F E R E N C E S
I Albright, E.C. & Larson, F.C. Metabolism of L-thyroxine by human tissue slices. J . clin. Inwst. 38, 1899, 1959.
2 Alexander, N.M. & Jennings, J.F. Analysis for total serum thyroxine by equilibrium competitive protein hinding on small re-usable sephadex columns. Clin. Chem. 20, 553, 1974.
13 Gaitan, J.E., Wahner, H.W., Gorman, C.A. & Jiang, N. Measurement of triiodothyronine in unextracted urine. J . Lab. d i n . Med. 86, 538, 1975.
14 Habermann, J., Horn, K. Ulbrecht, G. & Scriba, P.C. Simultane radioimmunologische Bestimmung von Thyroxin (T4) and Trijodthyronin (T3) in Urin. J . elin. Chem. clin. Biochem. 14, 595, 1976.
15 Hansen, H.H. Sephadex binding of '3'I-lahelled L-triiodothyronine as a test of thyroid function. Scand. J. clin. Lab. Invest. 18, 240, 1966.
16 Hufner, M. & Hesch, R.D. Triiodthyronine deter- minations in urine. Lancet, i , 101, 1973.
17 Ishihara, A., Loos, U. , Rothenbuchner, G.R. Pfeiffer, E.F. Thyroxine (T4) assay in urine as an index for the diagnosis of thyroid function. Acta Endocr. (Kbh.), 79, Suppl. 184, 79, 1974.
18 Kruse, V. Production and evaluation of high- quality thyroxine antisera for use in radioimmuno- assay. Scand. J . elin. Lab. Invest. 36, 95, 1976.
19 Larsen, P.R. Salicylate-induced increase in free triiodothyronine in human serum. J . d i n . Invest. 51, 1125, 1972.
20 Rogowski, P., Siersbaek-Nielsen, K. & Hansen, J.M. Age dependent changes in urinary excretion of triiodothyronine. Acta. Endocr. (Kbh.), Suppl. 199, 342, 1975.
21 Rogowski, P. & Siersbzk-Nielsen, K. Radio- immunoassay of thyroxine and triiodothyronine in urine using extraction and separation on Sepha- dex@ columns. Scand. J. elin. Lob. Invest. 37, 729, 1977.
22 Shakespear, R.A. & Burke, C.W. Triiodothyronine and thyroxine in urine. I . Measurements and application. J. elin. Endocr. 42, 494, 1976.
23 Weeke, J. & Gundersen, H.J. Correlation between short-time and diurnal variations in serum TSH and free thyroid hormones in normal man. Acta Endocr. (Kbh.), 82, Suppl. 204, 11, 1976.
