|
|
||||||||
CLINICAL STUDIES |
Department of Nuclear Medicine, Hôpital Salengro, CHRU, 59037 Lille Cedex, France1 Department of Endocrinology and Metabolic Diseases, CHU Rangueil, Toulouse, France2 Clinic of Endocrinology,, CHRU, Lille, France3 Service de Médecine interne,, Hôpital de Hautepierre, Strasbourg, France4 Institut de Physique Biologique,, Strasbourg, France5 Clinic of Endocrinology,, CHU, Montpellier, France and 6 Department of Endocrine Surgery,, CHRU, Lille, France
(Correspondence should be addressed to M d'Herbomez; Email: m-dherbomez{at}chru-lille.fr)
| Abstract |
|---|
|
|
|---|
Design: A prospective multicenter study.
Patients: Three hundred and seventy-five clinically euthyroid subjects.
Methods: We used five different CT immunoassays. Sera were assayed for the concentration of TSH, gastrin, procalcitonin, urea, calcium, and anti-thyroperoxidase antibodies.
Results: Screening for the various potential causes of hypercalcitoninemia led to the exclusion of 23% of the sera. Our reference value analysis dealt with 287 subjects (142 men and 145 women). The proportion of samples in which no CT was detected varied from 56% (for assay D) to 88% (for assay C). We observed significant correlations (whose magnitude depended on the assay used) between CT levels and age or body mass index (BMI) (primarily in men). The distribution of CT levels showed that 4.7, 9.8, 2.5, 6.5, and 8.0% of the values were over 10 pg/ml respectively. These values corresponded essentially to samples from 11 male subjects (median age: 55 years), most of whom were smokers. The highest CT values were around twice as high in men than women, and were higher in smokers than non-smokers.
Conclusion: In clinical practice (and after having excluded the usual causes of raised CT levels), the interpretation of CT assay results must take into account i) the method used; ii) the patient's gender, age, and weight; and iii) the potential influence of cigarette smoking.
| Introduction |
|---|
|
|
|---|
However, CT is not a specific marker for MTC because raised serum CT levels can equally be observed in other circumstances, including C-cell hyperplasia (CCH) (14, 15), other endocrine tumors (pulmonary and pancreatic tumors) (16), kidney failure (17, 18) and, more rarely, in auto-immune thyroid disease (19, 20), hypergastrinemia (21), sepsis (22), type 1A pseudohypoparathyroidism (23), and in the presence of heterophilic antibodies (24). The prevalence of hypercalcitoninemia not due to MTC varies from 0.3 to 4.5% (2, 3, 4, 5, 6, 7, 8, 9). The assay's high cost and lack of specificity (25) has dissuaded health authorities in the USA from adopting systematic testing (1, 26).
In contrast to other biological parameters in thyroid function testing (thyroid-stimulating hormone (TSH) and/or anti-thyroid antibodies), there are no published criteria for the selection of control subjects in order to establish CT reference values (27). Certain cases of hypercalcitoninemia could be due to inappropriate CT reference values because the latter depend on the assay type and indeed, for a given assay, the country in which it is used (3, 4, 28, 29).
The aim of the present study was to re-evaluate CT reference values by introducing criteria for the selection of control subjects, envisaging the use of different reference values for men and women respectively, and looking for a possible correlation between CT levels and subject's age.
| Materials and methods |
|---|
|
|
|---|
In this prospective multicenter study, 375 sera were sampled in subjects with no medical history of thyroid disease. No goiter was apparent and neck palpation did not reveal any abnormalities. Any subject was taking medication that could have interfered with thyroid function testing (with the exception of estrogens). Sera were aliquoted and stored at –20 °C prior to analysis.
Biological euthyroidism was confirmed by a TSH assay with a functional sensitivity <0.01 mUI/l (TSH architect from Abbott Diagnostics), for which the reference values range from 0.3 to 3.6 mUI/l. In order to eliminate other potential causes of raised CT, we determined calcemia and uremia at all the investigating centers. Tests for gastrin (GASK-PR from CisBio International, Gif-sur-Yvette, France; normal range 30 to 120 pg/ml), TSH, procalcitonin (sensitive ProCT on Kryptor from BRAHMS, Berlin, Germany), CT, and anti-thyroperoxidase antibodies (TPOAb from BRAHMS; positive if >60 UI/l) were all performed in the same laboratory.
CT assays
Five different CT immunoassays (based on radioisotopic, enzymatic, or luminescent labels) were tested: assay A, immunoradiometric assay-human calcitonin (IRMA-hCT) (CisBio International); assay B, CT on Advantage (Nichols Institute, San Clemente, CA, USA); assay C, CT on Immulite 2000 (DPC, La Garenne-Colombes, France); assay D, Calcitonin-ELISA 7024 (Biomerica, Newport Beach, CA, USA); and assay E, CT-USA-IRMA, KIP0429 (Biosource, Nivelles, Belgium). All recognize the mature form of CT, and the main characteristics of each kit are given in Table 1. The assays were performed as single runs, according to the respective manufacturer's recommendations. Two of the assays (B and C) are automated. The inter-assay coefficients of variation were <10% for all measured CT levels with the automated assays (B and C), and <10% only for the CT concentrations higher than 10 pg/ml with the manual assays (A, D, and E).
|
| Results |
|---|
|
|
|---|
Eighty-eight sera (23%) were eliminated due to abnormal TSH levels (n=12), and/or the presence of anti-thyroperoxidase antibodies TPOAb (n=34) and/or high gastrinemia, uremia, and calcemia values (in 13, 40, and 2 cases respectively). Several sera were eliminated for at least two reasons. All procalcitonin levels were in the normal range. The data reveal that auto-immune status does not greatly interfere with CT levels because only 1 of the 34 patients (i.e. 3%) presented a CT level >10 pg/ml according to all the five methods. In contrast, high urea levels were associated with a higher frequency of CT levels >10 pg/ml. Depending on the assay used, the proportion ranged from 5% (in assay C) to 23% (assay B) of the patients.
A 61 year old male subject undergoing post-operative and post-radiotherapy follow-up for a non-secreting pituitary macroadenoma (with no known thyroid disease) presented high CT levels, ranging from 313 to 473 pg/ml according to the various assays. The neck palpation was negative. The patient underwent thyroid echography (revealing a 1 cm thyroid nodule on the left lobe) and then a total thyroidectomy with dissection of the central and left jugular-carotid lymph nodes. Histological examination confirmed the presence of an 8 mm MTC in the left lobe, whereas the right thyroid lobe was normal. We did not observe recurrent or left jugular-carotid metastatic adenopathy. A negative test for the rearranged during transfection (RET) proto-oncogene mutation indicated that this patient presented a sporadic MTC. The post-operative baseline serum CT level was <3 pg/ml.
Reference ranges of CT levels
Hence, our CT reference value analysis dealt with 287 sera (142 men and 145 women; median age: 45 years (range 17–83 years)). Among the men, 45 were active smokers and 3 were former smokers (i.e., 33.8%), whereas 27 women were smokers (i.e., 18.6%). The CT concentration distributions established with the five assays are shown in Figs 1 and 2A and B. The indicated detection thresholds correspond to the functional sensitivity concentrations stated by the manufacturers. The number of undetectable results varied according to the detection threshold of each method and ranged from 154 (56%, method D) to 244 (88%, method C). For three assays, it was possible to dissociate the reference values for men and for women with sufficient precision. The maximum concentrations observed in men were twice those recorded in women.
|
|
Interference with cigarette smoking
After excluding 88 sera, 11 sera from male subjects (median age: 55 years) presented concordant CT levels of >10 pg/ml in at least three different assays (Fig. 1, assay A in the oval line). All the latter subjects were active smokers, except for one former smoker and one who had never smoked. The maximum CT concentration observed was 21.4 pg/ml.
Our data analysis enabled us to separate the values obtained in men from those found in women and to take into account the subject's smoking status. We report the maximum values observed for each assay in Fig. 3. For three assays (A, B, and C), positive smoking status significantly increased the interpretation threshold. This was not true with assays D and E. In addition, we calculated the median CT values for the non-smoker and smoker females and males with each assay. We found higher median CT concentrations for the smoker versus non-smoker males. The differences were significant with the assays A (P=0.018), B (P=0.001), C (P=0.012), D (P=0.036), and E (P=0.035). No significant differences were noted with females.
|
| Discussion |
|---|
|
|
|---|
Twenty-three percent of the sera were excluded from our analysis. The main reasons for elimination were marked uremia (accounting for 10% of the excluded sera) and the presence of TPOAbs (9.1%). Raised CT levels have been described (16, 17, 18) in cases of chronic kidney failure and in hemodialyzed patients (in up to 25% of the latter subjects). The mechanism underlying this elevation has not been well established. There are two possible hypotheses – either a modification of clearance or hyper-production – because a correlation between CT levels and the intensity of glomerular filtration has not been observed.
The prevalence of MTC is low, and among 1000 patients presenting nodular thyroid disease, only three to five patients will have MTC (2, 3, 9). On the basis of our 375 sera tested, we detected one case of sporadic MTC in a male subject. Systematic use of a CT assay enabled optimal care provision for this MTC patient, apparently with complete recovery.
Our reference value study dealt with 287 subjects (142 men and 145 women), i.e., in compliance with a minimum sample size of 120 recommended by Baloch et al. (27). The proportion of smokers was higher for the male subjects (33.8%) than the female subjects (18.6%).
In Figs 1 and 2A and B, we show the distribution of CT concentrations determined using the five assay kits. It is clearly apparent that a non-negligible proportion of the values is situated above 10 pg/ml (4.7, 9.8, 2.5, 6.5, and 8.0% for assays A, B, C, D, and E respectively). The maximum observed concentration was 28.1 pg/ml (for assay D, which often yielded isolated high values that were not confirmed by the other assays). The cut-off of 10 pg/ml recommended as the normal threshold by the assays A, C, and E (and even 12 pg/ml for B in men) appears to be too low. Our data are more in agreement with the threshold of 20 pg/ml chosen by Pacini's and Elisei's groups in Italy (3, 13). The recent study by Costante et al. (32) on 5817 assays indicated a prevalence of hypercalcitoninemia (>10 pg/ml) of 4.87%, including 0.26% for medullary cancers and 0.12% for CCH. It is noteworthy that the great majority of these cases of hypercalcitoninemia (77%) are due to CT levels between 10 and 20 pg/ml. Should we have different reference values for men and for women? Two of the five kits tested suggest doing so in their reference standards. A postmortem study (33) has shown that physiologically men have twice as many C-cells as women. Precise measurement of CT levels in women requires the use of techniques with low functional sensitivity and a measuring range that is well suited to low concentrations (30, 31, 34). The dissociation of CT levels in men and women also yields a ratio of 2 (i.e., as for the C-cell numbers). A single normal threshold of 20 pg/ml appears to be too high, in view of the values observed in female non-smokers. Normal more specific CT ranges (males versus females and smokers versus non-smokers) will permit the reduction of the gray zone of suspiciously elevated CT levels and possibly ameliorate the positive predictive value of basal CT for the diagnosis of medullary thyroid carcinoma.
We observed significant correlations between CT levels and patient age, the magnitude of which depended on the assay technique used and the population tested. The correlation coefficients were low (around 0.25 in men and half of this value in the overall population) and were more significant in men than women. Verga et al. (28) did not detect a correlation between CT levels (measured using assay B) and age in a population of 125 children and 98 male and female adults. This discrepancy can perhaps be explained by the fact that higher (and thus more easily detectable) values are found in men. The high proportion of undetectable values in women may bias this approach, and the same is true for correlations between the CT concentration and the BMI. Again, the significant correlations essentially concerned our male subjects.
Our in-depth analysis (presented in Table 2) identified 12 sera that corresponded to moderate hypercalcitoninemia. Of the 12 sera samples all were from male subjects (median age: 55), ten of whom were active smokers, one former smoker, and one subject who had never smoked. Hence, cigarette smoking may be a newly identified cause of elevated CT levels. This factor has not been taken into account by the previous studies in this field. We found significant differences between the median CT concentrations for only non-smoker and smoker males, because the number of smokers was higher in males than females, and because the observed CT concentrations were so elevated and measured with a better accuracy. The assays A, B, and C gave the better significance values. The assay D gave non-specific effects (30, 31). The action of tobacco on thyroid follicles is quite well known (35, 36, 37, 38). Tobacco use constitutes a risk factor for goiter and its effect is accentuated in the presence of iodine deficiency. Tobacco consumption has contrasting effects on thyroid, i.e., both stimulatory and inhibitory influences (33). The involvement of a variety of mechanisms of action has been evoked: the main toxins cited are thiocyanate (which inhibits iodide transport and organification) and 2,3-OH pyridine (which inhibits thyroxin deiodination). Cigarette smoking can aggravate tissue hypoxia and exert significant immunomodulatory effects (38), and the reversibility of cessation of tobacco consumption is not yet well understood (36). C-cells are distributed uniformly throughout the organism; they can be found not only in the thyroid and the thymus but also in the liver, lungs, duodenum, and jejunum. The effect of tobacco on pulmonary neuroendocrine cells has received more attention (39), and it has been established that smoking can increase the number of neuroendocrine cells and the secretion of peptides such as CT. With respect to thyroid C-cells, a single study on hamsters by Tabassian et al. (40) has revealed a CT elevation that may be related to the nicotine content of cigarettes. In order to determine the precise mechanisms of tobacco action on thyroid C-cells further studies would be required. Vantyghem et al. (41) observed an increase in CT concentrations in patients with high chronic alcohol and tobacco consumption and reported that CT levels did not normalize after the withdrawal of alcohol, suggesting that tobacco may play a predominant role.
|
In conclusion, this clinical study shows that all CT immunoassays are not equivalent. The most accurate assays allow dissociation of the CT normal ranges for males and females. The clinical and biological characterization of our control subjects revealed that men and women differ in terms of the maximum observed CT concentrations. Calcitonin levels are correlated with age and BMI (especially in men). Cigarette smoking can increase the concentration of CT, and so taking this factor into account will help interpret moderately elevated CT levels, especially in men.
| Acknowledgements |
|---|
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |