| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Endocrinology and Metabolic Diseases, Leiden University Medical Center, 2300 RC Leiden, The Netherlands
Address all correspondence and requests for reprints to: Jan W. A. Smit, M.D., Ph.D., Department of Endocrinology, C4-R, Leiden University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands. E-mail: jwasmit{at}lumc.nl.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Aims: The aim of our study was to investigate survival and specific death causes in a homogeneous cohort of DTC patients.
Patients: Patients included 366 consecutive patients with DTC who had all been treated according to the same protocol for initial therapy and follow-up.
Methods: Prognostic factors for DTC-related death were analyzed by univariate Cox regression analysis, followed by stepwise multivariate Cox regression analysis. Standardized mortality rates (SMR) were calculated using normal mortality rates for the entire Dutch population.
Results: During follow-up of 8.3 ± 4.6 yr, 82 patients (22.4%) died. At multivariate Cox-regression analysis, tumor stage T4, distant metastases, and advanced age were associated with an increased relative risk for DTC-related death. SMR for the entire group was 2.32. This could be explained by increased SMR in patients with stage T4, distant metastases, or advanced age. Death causes could be verified in 80 patients: 52 died of DTC, 28 due to other causes. Ten of the 20 patients with stage T1–3M0 died from thyroid carcinoma.
Conclusion: Relative risk for thyroid cancer-related death and SMR are significantly increased in patients with stage T4 and M1 or advanced age. Although death risk is not increased in T1–3 M0 patients, DTC contributed significantly to mortality in all patient categories.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
). Both approaches are susceptible to bias because referral to large centers may be subject to selection bias, whereas in population-based studies, initial treatment and therapies during follow-up may differ. Remarkably, initial therapy appeared to be standardized in only two of the 28 publications incorporated in Table 1 (8, 27).
|
We therefore studied survival in an unselected homogeneous group of 366 consecutive DTC patients who were treated according to the same protocol for initial therapy and follow-up. Prognostic indicators for mortality were analyzed by univariate and multivariate Cox regression analyses. Standardized mortality rates were calculated by comparing observed deaths with expected mortality in age- and gender-matched cohorts from the entire Dutch population.
In addition, we also verified specific death causes in all patients who died.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
With the exception of patients with unifocal T1N0M0 tumors, all patients received near-total thyroidectomy, followed by routine radioiodine ablative therapy with 2800 MBq I-131. In case of incomplete tumor resection or when metastases were present, 6000 MBq were administered after thyroidectomy.
Lymph node surgery was performed as follows: when lymph node metastases were the presenting symptom, a modified radical neck dissection (removal of lateral lymph nodes with preservation of sternocleidomastoid muscle, internal jugular vein, and accessory nerve) was performed at the time of total thyroidectomy. When lymph node metastases were not the presenting symptom, neck inspection was performed during thyroidectomy and suspected lymph nodes removed. When lymph node metastases became apparent during follow up, a modified radical neck dissection was performed.
Follow-up was performed according to a standard protocol, involving measurements of serum thyroglobulin levels both after thyroxin withdrawal (6 months after initial therapy) and during thyroxin therapy. Thyroxin therapy was aimed at TSH levels less than 0.1 mU/liter during 15 yr of follow-up. In addition, diagnostic 185 MBq I-131 scintigraphy after thyroxin withdrawal was performed 6 months after initial therapy.
In case of recurrent disease or metastases, surgery was attempted if the lesion was solitary and accessible, followed by additional radioiodine therapy (6000 MBq). If the tumor could not be removed surgically, radioiodine therapy was given and repeated if necessary. All radioiodine therapies were followed after 7 d by whole-body scintigraphy.
The follow-up period was from January 1986 to January 2003. The following data were documented: age at diagnosis, sex, date of diagnosis, histology, TNM (tumor-node-metastasis) stage, date of recurrence (if any), site of recurrence, thyroglobulin levels, and date of last follow-up or death. TNM stage was registered according to the fifth edition (29). We used the following end points of follow-up: date of death (82 patients), date of emigration (12 patients), and date of most recent contact (272 patients).
Death causes were analyzed in the 82 patients who had died during follow-up. Death causes were investigated using medical records, death certificates, and autopsy records. The cases were classified in five categories: 1) the patient definitely died from thyroid carcinoma (progressive metastatic or relapsing thyroid carcinoma without other potentially fatal disease and thyroid carcinoma mentioned as death cause in death certificate or autopsy file); 2) the patient most likely died from thyroid carcinoma (progressive metastatic or relapsing thyroid carcinoma without other potentially fatal disease in the last 3 months before death but thyroid carcinoma not specifically mentioned as death cause in death certificate); 3) the patient definitely died from another cause (documented fatal disease other than thyroid carcinoma, death cause other than thyroid carcinoma in death certificate or autopsy file, no progressive thyroid carcinoma at time of death); 4) the patient most likely died from another cause (documented fatal disease other than thyroid carcinoma without progressive thyroid carcinoma in the last 3 months before death; however, this fatal disease was not mentioned specifically as death cause in the death certificate); and 5) death cause unknown (in some cases, it was not possible to determine with certainty one single specific cause of death).
Statistical analysis
All data were expressed as mean ± SD, proportions, or absolute numbers. Survival curves were generated using Kaplan Meyer analyses. Prognostic factors, expressed as relative risks (RR), were analyzed by univariate Cox regression analysis. Indicators that were identified as significant for survival in univariate analysis were entered into a stepwise multivariate Cox regression analysis model.
Standardized mortality rates (SMRs) were calculated as follows: normal mortality rates for the Dutch population were obtained from the Dutch Central Bureau of Statistics (The Netherlands; www.cbs.nl) using mortality rates per sex, age groups of 5 yr, and four calendar periods (1986, 1990, 1995, 2000). Expected mortality rates were determined based on the person-years follow-up for each sex and age group and calendar period and compared with the observed mortality rate.
For analysis of death causes, comparison of proportions was performed by 
2 analysis or binomial logistic regression. Normally distributed data were compared by unpaired t testing. P < 0.05 was considered significant. SPSS for Windows (version 12.0; SPSS Inc., Chicago, IL) was used to perform data analyses.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Three hundred sixty-six consecutive patients were included, 91 male and 275 female patients. Mean age at the time of diagnosis was 47.6 ± 18.0 yr. Mean follow-up was 8.3 ± 4.6 yr (range 1–16 yr). A total of 3050 yr of follow-up was available. Data for the survival analysis are given in Tables 2 and 3 and Fig. 1.
|
|
|
End points of follow-up were defined as follows: death (n = 82), date of most recent contact (n = 272), and date of emigration (n = 12), respectively. Eighty-two patients died during follow-up (2 months to 16 yr after diagnosis), and mean age at death was 67.2 ± 11.5 yr (range 30–93 yr). Total death rates and thyroid carcinoma-specific death rates are given in Table 2
and Fig. 1B. The 10-yr total and cause-specific survivals were 76.8 and 84.9%, respectively (Fig. 1).
Prognostic indicators for thyroid cancer-related death were identified by univariate and multivariate Cox-regression analysis (Table 2). The following parameters were associated with a significantly increased RR for thyroid cancer-related death mortality risk by univariate analysis: follicular histology (RR 1.99), stage T4 (RR 11.46), presence of lymph nodes (RR 2.01), presence of distant metastases (9.69), and advanced age (1.08/yr). Gender did not contribute to death risk. In multivariate Cox regression analyses, stage T4, the presence of distant metastases, and age were associated with an increased RR for thyroid cancer-related death.
Standardized mortality rates were calculated by comparing total death rate in the patient group (observed deaths) and expected mortality in age- and gender-matched cohorts of the entire Dutch population. The SMR for the entire group was 2.32. SMR was also calculated for patients categorized according to prognostic indicators that were significant at multivariate Cox regression analyses: SMR in T1–3M0 patients was 1.05 [confidence interval (CI) 0.64–1.56], T4 M0 4.28 (2.79–6.09), and M1 3.81 (2.58–5.28). SMR was also significantly influenced by age: in patients younger than 55 yr at the time of diagnosis, SMR was not significantly increased, in contrast with patients with age 55 yr or older.
Fifty patients had recurrent disease (six local recurrent disease, 14 lymph node metastases, 30 distant metastases). The mean interval from diagnosis to recurrence was 4.6 ± 0.5 yr. Eight of these patients were cured and 22 died. At the end of follow-up, 252 patients were alive without tumor and 32 with tumor. Forty-eight patients had serum thyroglobulin levels greater than 1 µg/liter at the end of follow-up. Of these patients, 32 had tumor, whereas in 16 no tumor was detectable.
Death cause analysis
In two patients, no specific death cause could be established. Therefore, these two patients were not included in the analyses. Specific death causes of the remaining 80 patients are given in Tables 4 and 5. In general, 65% of the deaths were caused by thyroid carcinoma. TNM stage significantly influenced the likelihood to die from thyroid cancer, compared with other death causes (P = 0.005). Further analyses revealed that this was predominantly related to the presence of distant metastases. However, in 10 of the 20 patients with stage T1–3M0, the specific cause of death was related to thyroid carcinoma (Table 4). In other words, although stage T1–3M0 is not associated with an increased risk for death (Table 2), the tumor contributed to the death cause in this category of patients. The likelihood that thyroid cancer is the death cause in T1–3M0 patients (0.50, CI 0.13–0.77) is comparable with that of patients with stages T4 (0.52, CI 0.21–0.75) and M1 (0.87, CI 0.71–0.97). Apparently, the T1–3M0 subgroup of tumors also contains tumors with an unfavorable prognosis. Histology, gender, and age at diagnosis were not significant predictors for specific death cause.
|
|
As expected, the time interval from diagnosis to death differed significantly between tumor stages in patients who died from thyroid carcinoma [7.4 ± 5.4 yr in T1–3M0 patients vs. 3.3 ± 2.5 yr in patients with stage M1 (P = 0.009)] but not in patients who died from other causes.
Analysis of the other death causes revealed a remarkably high percentage of other malignancies (57% of death causes), whereas cardiovascular disease was underrepresented (18%).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
All prognostic factors are strongly interrelated, and, i.e. large tumors, distant metastases, poorly differentiated histology, and absence of radioiodine uptake, are more frequently observed in older patients. We therefore performed univariate and multivariate Cox regression analysis to identify independent prognostic indicators. At multivariate Cox regression analyses, we identified stage T4, the presence of distant metastases, and age as independent indicators associated with an increased relative risk for thyroid carcinoma-related death. This is entirely in line with the only other publication, in which a comparable protocol for initial therapy was used as in our clinic (8).
Stages T1–3, gender, histology, and lymph node metastases were not associated with an increased relative risk for thyroid cancer-related death at multivariate analysis. The fact that lymph node stage was no independent predictor for death is in line with many other studies (Table 1) but is also a subject for ongoing debate. The fact that in our patient group all patients underwent initial radioiodine ablation therapy combined with an active surgical approach in case of the presence of lymph node metastases may explain the fact that lymph nodes were not associated with an increased death risk. It has to be noted, however, that initial lymph node staging was based on palpation during surgery and postablative whole-body scintigraphy and that the conclusions with regard to lymph nodes should be restricted to this staging procedure.
Total mortality of the entire patient group was increased in comparison with expected mortality in age- and gender-adjusted cohorts of the entire Dutch population. This increased standardized mortality rate was present in patient categories with T4, M1, or advanced age but not in patients with stage T1-3M0.
The two methods of mortality analysis (multivariate Cox regression and SMR) are complementary: univariate and multivariate Cox regression analyses provide significant indicators for survival within the study group. SMR represent death risks in comparison with a control group.
The method of SMR has several limitations. First, the disease that is subject of study has to have a very low prevalence in the general population, which is true for DTC. Second, excess deaths in the patient group may not be causatively related to the disease but may also be caused by associated diseases or toxicity of treatment (radioiodine, TSH suppression). Third, expected deaths in the patient group may be lower than the control group because these patients are under medical surveillance, revealing other health hazards at an earlier stage.
An interesting finding is that in 10 of the 20 patients with stage T1–3M0, the specific cause of death was related to thyroid carcinoma, although this stage is not associated with an increased death risk. Conversely, the number of non-thyroid cancer-related deaths (10) in T1–3M0 patients is lower than the expected number of deaths (19.92) in age- and sex-matched cohorts of the general population (SMR 0.50, CI 0.23–0.87). Three explanations for the lower SMR for non-thyroid cancer deaths can be proposed. First, a patient who dies from DTC apparently does not reach the end point of non-thyroid cancer-related death, even if he has the same risk profile as patients without DTC. Second, from a theoretical point of view, it can be proposed that DTC or its treatment protects against certain fatal diseases. Third, intense medical care may reveal potential fatal diseases at an early stage, making cure possible.
We believe that within tumor stage T1–3M0, there is a subgroup with unfavorable tumor characteristics that is not identified by clinical staging and/or conventional histological characterization. The assumption that the distribution of death causes in patients who do not have an increased SMR for death represents the pattern of death causes in the general population may therefore not be true. We realize, however, that with a longer follow up, the distribution of death causes in the T1–3M0 group may be expected to shift toward nontumor-related death causes. Nonetheless, we believe that all patients with DTC who have received total thyroidectomy and radioiodine ablation should undergo a TSH-stimulated thyroglobulin measurement at least once after initial therapy and should be followed up by yearly thyroglobulin measurements on thyroxin.
Another finding was the remarkably high percentage of other malignancies as non-thyroid cancer-related death cause (58%). In the study of Akslen et al. (18), the percentage of other malignancies as death cause was 38%. This finding points to the importance of research into associations of thyroid carcinoma with other malignancies. This has been documented in some tumor syndromes, like the Cowden syndrome or the familial adenomatous polyposis coli syndrome (30). In addition, the question whether the initial therapy of thyroid carcinoma (e.g. radioiodine therapy) may contribute to the susceptibility to other malignancies has not been solved conclusively, but the short interval between initial therapy and death makes it unlikely that there is a causative role in this series. In addition, the fact that our patients are under medical control may reveal other tumors at an earlier stage.
In conclusion, we found that RR for thyroid cancer-related death and SMR are increased in patients with an initial stage T4 or M1 and advanced age. However, thyroid carcinoma contributed substantially to mortality in all patients, including patients with stage T1-3M0.
|
Footnotes |
|---|
First Published Online November 1, 2005
Abbreviations: CI, Confidence interval; DTC, differentiated thyroid carcinoma; RR, relative risk; SMR, standardized mortality rate; TNM, tumor-node-metastasis.
Received June 14, 2005.
Accepted October 26, 2005.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Sciuto, L. Romano, S. Rea, F. Marandino, I. Sperduti, and C. L. Maini Natural history and clinical outcome of differentiated thyroid carcinoma: a retrospective analysis of 1503 patients treated at a single institution Ann. Onc., October 1, 2009; 20(10): 1728 - 1735. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. A Verburg, U. Mader, M. Luster, and C. Reiners Histology does not influence prognosis in differentiated thyroid carcinoma when accounting for age, tumour diameter, invasive growth and metastases Eur. J. Endocrinol., April 1, 2009; 160(4): 619 - 624. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Roti, E. C degli Uberti, M. Bondanelli, and L. E Braverman Thyroid papillary microcarcinoma: a descriptive and meta-analysis study Eur. J. Endocrinol., December 1, 2008; 159(6): 659 - 673. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. T. Kloos Approach to the Patient with a Positive Serum Thyroglobulin and a Negative Radioiodine Scan after Initial Therapy for Differentiated Thyroid Cancer J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1519 - 1525. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Brabant Thyrotropin Suppressive Therapy in Thyroid Carcinoma: What Are the Targets? J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1167 - 1169. [Full Text] [PDF]
|
|
|
|
 |
|
 S.-H. Ahn, Y. Henderson, Y. Kang, C. Chattopadhyay, P. Holton, M. Wang, K. Briggs, and G. L. Clayman An Orthotopic Model of Papillary Thyroid Carcinoma in Athymic Nude Mice Arch Otolaryngol Head Neck Surg, February 1, 2008; 134(2): 190 - 197. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Y. Farrag, F. R. Lin, C. W. Cummings, J. J. Sciubba, W. M. Koch, P. W. Flint, and R. P. Tufano Importance of routine evaluation of the thyroid gland prior to open partial laryngectomy. Arch Otolaryngol Head Neck Surg, October 1, 2006; 132(10): 1047 - 1051. [Abstract] [Full Text] [PDF]
|
|
Read all eLetters
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
