ORIGINAL
ARTICLE
E n d o c r i n e
C a r e
Thyroid Function Tests and Mortality in Aged Hospitalized Patients: A 7-Year Prospective Observational Study Pedro Iglesias, Elena Ridruejo, Angélica Muñoz, Florentino Prado, María Cruz Macías, María Teresa Guerrero, Pilar Tajada, Carmen García-Arévalo, and Juan José Díez Department of Endocrinology (P.I., J.J.D.), Hospital Ramón y Cajal, 28034 Madrid, Spain; and Departments of Geriatrics (E.R., A.M., F.P., M.C.M., M.T.G.) and Biochemistry (P.T., C.G.-A.), Hospital General, 28007 Segovia, Spain
Context: Several alterations in thyroid function test (TFT) results have been associated with mortality in elderly patients. Objective: Our aim was to investigate the relationship between TFT results and all-cause and cardiovascular (CV) mortality in aged hospitalized patients. Design: A 7-year prospective observational study was conducted. TFTs were performed at hospital admission, and mortality was registered in the follow-up period. Patients: Participants were 404 patients aged ⬎65 years admitted to the Department of Geriatrics, Hospital General, Segovia, Spain, for any reason during 2005. Main Outcome Measures: The study evaluated the association between TFT results and mortality from all causes and CV diseases. Methods: TSH, free T4, and free T3 (FT3) were measured on the first day of admission. In-hospital and total survival times, number of deaths, and all-cause and CV mortality were registered until the census date on January 1, 2012. Results: During the study, 323 patients (80%) died. Kaplan-Meier analysis showed that median survival time for all-cause mortality was significantly lower in patients in the first tertile of serum FT3, in the first tertile of TSH, and in the first tertile of serum free T4 concentrations. Multivariate adjusted Cox regression analysis showed that the history of cancer (hazard ratio, 1.60; 95% confidence interval, 1.12–2.28; P ⫽ .009), age (1.03; 1.01–1.06; P ⫽ .003), and FT3 levels (0.72; 0.63– 0.84; P ⬍ .001) were significant factors related to all-cause mortality. The cause of death was known in 202 patients. Of this group, 61 patients (30.2%) died of CV disease. Patients in the first tertile of TSH and FT3 exhibited a significant higher mortality due to CV disease. In the adjusted Cox regression analysis, FT3 was a significant predictor of CV mortality (0.76; 0.63– 0.91; P ⫽ .004). Conclusions: Alterations in TFT results during hospitalization are associated with long-term mortality in elderly patients. In particular, low FT3 levels are significantly related to all-cause and CV mortality. (J Clin Endocrinol Metab 98: 4683– 4690, 2013)
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2013 by The Endocrine Society Received November 7, 2012. Accepted September 25, 2013. First Published Online October 30, 2013
doi: 10.1210/jc.2012-3849
Abbreviations: CHF, congestive heart failure; CHD, coronary heart disease; CV, cardiovascular; FT3, free T3; FT4, free T4; NTIS, nonthyroidal illness syndrome; TFT, thyroid function test.
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
jcem.endojournals.org
4683
4684
Iglesias et al
Thyroid and Mortality in the Elderly
hyroid dysfunction is a common disorder in the general population, mainly in women of advanced age (1–5). The prevalence of overt hypothyroidism and hyperthyroidism ranges between 1% and 7% and between 0.5% and 3.0%, respectively (2, 3, 6 – 8), increasing to 14% to 18% and higher than 5% when subclinical hypothyroidism and subclinical hyperthyroidism are considered (3, 8). Untreated thyroid disorders in elderly individuals are associated with significant morbidity (9 –11). Several studies (12–16), but not all (10, 17), suggest a relationship between altered thyroid function test (TFT) results and mortality in elderly individuals. Low levels of TSH or elevated levels of free T4 (FT4) have been related to an increased mortality rate in this population (12–15, 18). On the contrary, a high TSH level or low serum FT4 level was associated with lower mortality (13, 15, 16). A recent survey showed that a single measurement of thyroid function (TSH and T4) was not able to predict total or cause-specific mortality in a cohort of community-dwelling older men over an 8-year period of follow-up (17). Alterations in TFT results have also been associated with morbidity and mortality in older hospitalized patients. A recent retrospective study has shown that low T4 and high TSH levels were associated with a worse prognosis in these patients (19). It was also reported that a low serum T3 concentration behaves as a powerful predictor
T
Table 1.
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
for mortality during hospitalization in elderly patients (20, 21). In summary, some epidemiological studies suggest that different derangements in TFT results are related to mortality in elderly individuals in diverse clinical situations. Our aim has been to analyze whether a single TFT (TSH, FT4, and free T3 [FT3]) measurement could be a predictive factor of a better or worse vital prognosis not only during hospitalization but also in the long-term, as well as its relationship with all-cause and cardiovascular (CV) mortality in elderly patients admitted because of an acute process.
Patients and Methods Study design A 7-year (between January 1, 2005, and January 1, 2012) prospective observational study including patients older than 65 years hospitalized because of an acute process during 2005 was carried out. Patients were recruited from a previous cross-sectional study (21). All patients admitted to our geriatric unit were included in the study regardless of the cause of hospitalization. All patients taking medications known to modify thyroid function (antithyroid drugs, thyroxine, glucocorticoids, octreotide, dopamine, lithium, or amiodarone) were excluded. Demographic, clinical, and laboratory data were recorded (Table 1). In-hospital and total survival times, number of deaths,
Clinical and Analytical Data and Prevalent Comorbidities of the Study Patients
No. of patients, % Age, y Hypertension, % Diabetes, % Dyslipidemia, % Cardiovascular disease, % CHD Stroke Peripheral arterial disease CHF Arrhythmia Cancer, % Alterations in TFT results, % NTIS Subclinical hypothyroidism Overt hypothyroidism Subclinical hyperthyroidism Overt hyperthyroidism Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Thyrotropin, mU/L FT4, pmol/L FT3, pmol/L
Women
Men
Total
247 (61.1) 86.7 ⫾ 6.3 153 (61.9)a 67 (27.1) 37 (15.0) 84 (34.0) 18 (7.3) 46 (18.6) 8 (3.2) 45 (18.2) 94 (38.0) 26 (10.5) 180 (72.9) 151 (61.1) 16 (6.5) 2 (0.8) 5 (2.0) 6 (2.4) 27.4 ⫾ 6.2 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.6) 17.2 (15.2–19.7)a 3.7 ⫾ 0.9
157 (38.9) 84.8 ⫾ 6.3 67 (42.7) 34 (21.7) 23 (14.6) 46 (29.3) 20 (12.7) 19 (12.1) 18 (11.5) 16 (10.2) 52 (33.1) 28 (17.8) 115 (73.2) 100 (63.7) 7 (4.4) 2 (1.3) 3 (1.9) 3 (1.9) 27.2 ⫾ 3.5 130 (110 –140) 70 (60 – 80) 1.5 (0.8 –2.5) 16.3 (14.3–18.4) 3.5 ⫾ 0.9
404 85.9 ⫾ 6.4 220 (54.4) 101 (25.0) 60 (14.8) 130 (32.2) 38 (9.4) 65 (16.1) 26 (6.4) 61 (15.1) 146 (36.1) 54 (13.4) 295 (73.0) 251 (62.1) 23 (5.7) 4 (1.0) 8 (2.0) 9 (2.2) 27.3 ⫾ 5.1 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.5) 16.9 (14.8 –19.1) 3.6 ⫾ 0.9
Data represent the number of patients (percentage) and/or the mean ⫾ SD for normally distributed data and median (interquartile range) for nonnormally distributed data. a
Level of significance for multiple comparisons was adjusted using the Bonferroni correction to ␣: P ⫽ .004
doi: 10.1210/jc.2012-3849
and all-cause and CV mortality were registered until the census date on January 1, 2012. The vital status of every patient was evaluated through telephone, hospital medical reports, civil registration, and contact with the primary care physicians of the patients. The cause of death was obtained through the clinical information of the hospital medical reports and death certificates of the patients. CV death was defined as death from stroke, coronary heart disease (CHD), congestive heart failure (CHF), arrhythmia, myocardial infarction, or sudden death. This study was approved by our local ethics committee, and informed consent was given by subjects or their relatives before participation in the study.
Thyroid functional status Thyroid function was assessed by measuring serum concentrations of TSH, FT4, and FT3. Patients were classified according to the type and severity of thyroid dysfunction in overt hyperthyroidism (low TSH and high FT4 and/or high FT3), subclinical hyperthyroidism (low TSH and normal FT4 and FT3), overt hypothyroidism (high TSH and low FT4), and subclinical hypothyroidism (high TSH levels with normal FT4). Nonthyroidal illness syndrome (NTIS) or low T3 syndrome was also assessed.
Patients From an initial group of 447 patients, we excluded 43 patients who were taking thyroid-altering medications. We finally selected a group of 404 patients (Table 1). Clinical analytical data for the study patients are summarized in Table 1.
Hormone assays TSH, FT4, and FT3 serum concentrations were measured in all patients. Fasting samples of venous blood were obtained from an antecubital vein between 8:30 and 9:00 AM for hormonal quantifications on the first day of admission. Serum TSH, FT4, and FT3 were measured by an electrochemiluminescence immunoassay by using an E170 analyzer (Roche Diagnostics, Mannheim, Germany). Maximal intra- and interassay coefficients of variation as indicated in the package inserts of the commercial kits were 1.35% and 1.16% for TSH, 2.04% and 3.89% for FT4, and 2.87% and 10.17% for FT3. The sensitivities of the TSH, FT4, and FT3 assays were 0.005 mU/L, 0.3 pmol/L, and 0.4 pmol/L, respectively. Reference values were as follows: TSH, 0.4 to 5.0 mU/L; FT4, 11 to 23 pmol/L; and FT3, 3.9 to 6.8 pmol/L.
Statistical analysis For quantitative variables, results are expressed as mean ⫾ SD for normally distributed data and as median (interquartile range) for nonnormally distributed data. The normality of the data distribution was tested using the Kolmogorov-Smirnov test. Categorical variables are described as percentages. For comparisons of means between 2 groups of subjects, the Student t test was used for normally distributed data, and the Mann-Whitney test was used for nonnormally distributed data. For the comparison of ⬎2 independent groups, one-way ANOVA and Kruskal-Wallis tests were used for normally and nonnormally distributed data, respectively. For categorical comparisons, the 2 test or the Fisher exact test was used. Bonferroni correction to the ␣ level was applied when tests with multiple comparisons were performed. Survival time was estimated by the Kaplan-Meier method; the log rank test was used to compare arms. Unadjusted and stepwise multivariate Cox regression models were used to assess
jcem.endojournals.org
4685
the effects of several quantitative (age, TSH, FT4, and FT3) and qualitative (sex, hypertension, diabetes mellitus, dyslipidemia, CV disease, and cancer) variables on the risk of death. Hazard ratios (HRs) and their 95% confidence intervals (CIs) for allcause and CV mortality were estimated. Patients were classified according to tertiles of TSH, FT4, and FT3 levels obtained from 3 approximately equal groups based on the data analyzed. This analysis implies a possible bias or lack of generalization of these tertiles in other patient groups. Differences were considered significant when P ⬍ .05.
Results TFT results and morbidity TFT results were abnormal in most of the patients (73%); NTIS was the most common alteration (251 patients [62.1%]), followed by subclinical hypothyroidism (23 patients [5.7%]), overt hyperthyroidism (9 patients [2.2%]), subclinical hyperthyroidism (8 patients [2.0%]), and overt hypothyroidism (4 patients [1.0%]) (Table 1). Prevalent comorbidities are shown in Table 1. The presence of diabetes, dyslipidemia, CV disease, and cancer did not result in significantly different serum concentrations of TSH, FT4, and T3. Only hypertensive patients showed significantly higher serum FT4 levels than nonhypertensive subjects. The main causes for hospital admission were CHF (77 patients [19.1%]), stroke (64 patients [15.8%]), respiratory tract infection (54 patients [13.4%]), acute digestive hemorrhage (25 patients [6.2%]), CHD (23 patients [5.7%]), exacerbation of chronic obstructive pulmonary disease (23 patients [5.7%]), cancer (14 patients [3.5%]), sepsis (11 patients [2.7%]), urinary tract infection (11 patients [2.7%]), anemia (7 patients [1.7%]), syncope (7 patients [1.7%]), acute gastroenteritis (7 patients [1.7%]), confusion syndrome (6 patients [1.5%]), bradycardia (6 patients [1.5%]), cellulitis (5 patients [1.2%]), pancreatitis (3 patients [0.7%]), and other conditions (61 patients [15.1%]). No significant differences in TSH and FT4 were found among these groups. Patients with sepsis had the lowest FT3 levels (2.7 ⫾ 0.8 pmol/L), which were significantly different from those found in patients with stroke (3.9 ⫾ 1.0 pmol/L, P ⫽ .002) and CHD (3.8 ⫾ 0.9 pmol/L, P ⫽ .001). TFT results and mortality During the study, 323 patients (80%) died (median survival time, 9 months; interquartile range, 1–31 months; 61 patients [15.1%] died during hospitalization). Distribution of the patients according to the reference values of thyroid parameters in relation to vital status and time of death is shown in Table 2. The percentage of patients with low T3 levels was significantly higher in the group of pa-
4686
Iglesias et al
Table 2. Status
Thyroid and Mortality in the Elderly
Distribution of Patients According to Reference Values of Thyroid Parameters in Relation to Their Vital
Dead in the Hospital (n ⴝ 61)
TSH FT4 FT3
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
Dead After Discharge (n ⴝ 262)
Survivors (n ⴝ 81)
Low
Normal
High
Low
Normal
High
Low
Normal
High
P Value
7 (11.5) 3 (4.9) 52 (85.2)
52 (85.2) 57 (93.4) 9 (14.8)
2 (3.3) 1 (1.6) 0
24 (9.2) 7 (2.7) 170 (64.9)
221 (84.4) 238 (90.8) 91 (34.7)
17 (6.5) 17 (6.5) 1 (0.4)
1 (1.2) 3 (3.7) 37 (45.7)
72 (88.9) 74 (91.4) 44 (54.3)
8 (9.9) 4 (4.9) 0
.007 NS ⬍.001
Abbreviation: NS, not significant. Data represent number of patients (%). Normal range: TSH, 0.4 to 5.0 mU/L; FT4, 11 to 23 pmol/L, and FT3, 3.9 to 6.8 pmol/L.
tients who died in the hospital compared with those who did after discharge. Only 9 of 144 patients (6.2%) with normal FT3 levels died during hospitalization, making that finding highly predictive of survival. Patients who died were older and showed lower values of FT3 than patients who survived (Table 3). The presence of abnormal TFT results was positively associated with mortality (P ⫽ .001). Among patients who died, both TSH (1.2 [0.7–2.2] vs 1.6 [0.9 –2.6] mU/L, P ⫽ .047] and FT3 (3.2 ⫾ 0.8 vs 3.7 ⫾ 0.9 pmol/L, P ⬍ .001) were lower in those who died during hospitalization than among those who survived to hospital discharge. No significant differences in FT4 were found. Of the 323 patients who died, we only could find the cause of death in 202 patients (62.5%). CV disease was the main cause of death in 61 of these patients (30.2%). We did not find any significant difference in clinical and analytical data and TFT results among patients who died of CV disease and those who died of other causes. No differences were found in TSH, FT3, and FT4 among patients Table 3. Study
who died of CV disease during hospitalization and those who died after discharge. Kaplan-Meier analysis showed a median survival time for all-cause mortality of 3.0 (95% CI, 1.1– 4.9), 13.0 (6.1–19.8), and 19.0 (12.1–13.3) months for patients belonging to the first (FT3 ⱕ3.18 pmol/L), second (3.18 ⬎ FT3 ⱕ 3.96 pmol/L), and third (FT3 ⬎3.96 pmol/L) tertiles of FT3, respectively (P ⬍ .001) (Figure 1). In regard to TSH, median survival times for all-cause mortality were 3.0 (95% CI, 0.6 –5.4), 17.0 (9.9 –24.1), and 12.0 (5.7– 18.3) months for patients belonging to the first (TSH ⱕ1.06 mU/L), second (1.06 ⬎ TSH ⱕ 2.09 mU/L), and third (TSH ⬎ 2.09 mU/L) tertiles of TSH, respectively (P ⫽ .006) (Figure 1). In the same way, median survival times for all-cause mortality were 6.0 (3.6 – 8.4), 19.0 (12.8 – 25.1), and 11.0 (6.6 –15.4) months for patients belonging to the first (FT4 ⱕ 15.4 pmol/L), second (15.4 ⬎ FT4 ⱕ 18.2 pmol/L), and third (FT4 ⬎ 18.2 pmol/L) tertiles of FT4, respectively (P ⫽ .004) (Figure 1). Similar findings for TSH (log rank test 6.4, P ⫽ .041) and FT3 (log rank test
Clinical and Analytical Data of the Study Patients at Entry According to Their Vital Status at the End of the
Patients, n (%) Age, y Hypertension, % Diabetes, % Dyslipidemia, % CV disease, % CHD Stroke Peripheral arterial disease CHF Cancer, % Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Thyrotropin, mU/L FT4, pmol/L FT3, pmol/L
Alive
Dead
81 (20) 84.0 ⫾ 6.8 48 (59.3) 15 (18.5) 11 (13.6) 27 (33.3) 5 (6.2) 16 (19.7) 2 (2.5) 10 (12.3) 8 (9.9) 27.3 ⫾ 6.0 135 (120 –150) 75 (70 – 80) 1.8 (0.9 –2.8) 17.2 (14.9 –19.3) 4.0 ⫾ 0.9
323 (80) 86.4 ⫾ 6.2 172 (53.2) 86 (26.6) 49 (15.1) 103 (31.9) 33 (10.2) 49 (15.2) 24 (7.4) 51 (15.8) 46 (14.2) 27.4 ⫾ 5.0 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.4) 16.8 (14.8 –19.0) 3.5 ⫾ 0.9
P Value .002 NS NS NS NS
NS NS NS NS NS NS ⬍.001
Abbreviation: NS, not significant. Data represent the number of patients (%) or the mean ⫾ SD for normally distributed data and median (interquartile range) for nonnormally distributed data. a
Level of significance for multiple comparisons was adjusted using the Bonferroni correction to ␣: P ⫽ .004
doi: 10.1210/jc.2012-3849
jcem.endojournals.org
4687
Figure 1. Kaplan-Meier survival analysis for all-cause (left panels; n ⫽ 404) and CV mortality (right panels; n ⫽ 283) in elderly patients admitted for acute illness stratified according to tertiles of TSH, FT4, and FT3.
25.6, P ⬍ .001), were found when mortality due to CV disease was analyzed. Survival analysis in relation to tertiles of FT4 did not show significant differences in this group of patients (Figure 1).
In an unadjusted Cox regression model, only age (years) (HR, 1.03; 95% CI, 1.01–1.0; P ⫽ .001) and FT3 (picomoles per liter) (0.71; 0.62– 0.82; P ⬍ .001) were significantly related to all-cause mortality; whereas in the
4688
Iglesias et al
Thyroid and Mortality in the Elderly
multivariate (adjusted) analysis, a history of cancer (1.60; 1.12–2.28; P ⫽ .009), age (1.03; 1.01–1.06; P ⫽ .003), and FT3 levels (0.72; 0.63– 0.84; P ⬍ .001) were significant factors related to all-cause mortality. In regard to CV mortality, only age (1.05; 1.02–1.07; P ⬍ .001) and FT3 levels (0.71; 0.60 – 0.83; P ⬍ .001) were significant factors in the unadjusted analysis. In the multivariate (adjusted) analysis FT3 (0.76; 0.63– 0.91; P ⫽ .004) behaved as a significant factor associated with CV mortality along with cancer (1.64; 1.06 –2.55; P ⫽ .027) and age (years) (1.05; 1.02–1.08; P ⫽ .003) (Table 4).
Discussion Our results clearly show a significant relationship between TFT results and mortality in aged hospitalized patients not only during hospitalization but also long term after hospital discharge. Low serum levels of TSH, FT4, and, in particular, FT3 were associated with increased all-cause mortality. Low FT3 was a significant predictor of all-cause mortality both before and after hospital discharge. In addition, low TSH and low FT3 levels were associated with higher CV mortality. However, multivariate analysis showed that only FT3 and age were independent predictors for CV mortality in this population. Lastly, we could not find any significant relationship between thyroid status and comorbidity with the exception of increased FT4 levels in hypertensive patients and decreased FT3 levels in patients who were admitted for sepsis. Several studies have investigated the prevalence of thyroid dysfunction, as well as its relationship with morbidity and mortality in elderly populations (12, 13, 15–18, 20 – 25). Some of them were performed in an apparently
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
healthy elderly population (12, 13, 15–18), whereas others were performed in hospitalized (20 –24) or institutionalized (26) elderly patients. Our findings confirm that the prevalence of alterations in TFT results in elderly hospitalized patients is notably high, about three quarters of patients, mainly with NTIS, as reported previously (20, 21, 23). The association between alterations in TFT results and comorbidity has been well established (27). In elderly individuals, some studies found a relationship between subclinical thyroid hyperfunction and atrial fibrillation but not with other CV disorders (10, 28). Others relate subclinical hyperthyroidism (TSH ⬍0.1 mU/L) and subclinical hypothyroidism (TSH ⬎10 mUl) with an increase in the risk of incident CHF in elderly patients with known CV risk factors or previous CV disease (27, 29). Our study shows no significant relationship between TFT results and prevalent CV disease or known risk factors for CV disease, including diabetes mellitus and dyslipidemia. Only hypertensive patients showed higher FT4 levels, although within the normal range, with FT3 and TSH levels similar to those of nonhypertensive patients. This observation is in accordance with a recent report showing a positive correlation between FT4 and left ventricular mass in hypertensive euthyroid patients (30). Low T3 syndrome affects ⬃60% to 66% of elderly patients hospitalized for acute illness (19, 21, 23). This alteration is usually transient, serum T3 levels normalizing in about 30% a month after hospital discharge (31). Our survey did not find any association between the presence of low T3 syndrome and the prevalent comorbidity, suggesting that this alteration would be more closely related to the acute illness than to the type of chronic conditions
Table 4. Nonadjusted and Adjusted HRs (with 95% CIs) for the Development of All-Cause and CV Mortality in Patients Classified According to Several Clinical and Analytical (TFTs) Variables All-Cause Mortality Unadjusted Analysis
CV Mortality Multivariate (Adjusted) Analysis Unadjusted Analysis
Variable
HR (95% CI)
P Value
HR (95% CI)
Male sex Age, y Hypertension Diabetes mellitus Dyslipidemia CV disease Cancer TSH, mU/L FT4, pmol/L FT3, pmol/L
1.05 (0.84 –1.31) 1.03 (1.01–1.05) 1.19 (0.96 –1.50) 1.01 (0.78 –1.29) 1.14 (0.84 –1.55) 0.97 (0.76 –1.24) 0.75 (0.54 –1.03) 1.00 (0.97–1.04) 0.98 (0.95–1.01) 0.71 (0.62– 0.82)
NS .001 NS NS NS NS NS NS NS ⬍.001
NS .003 NS NS NS NS 1.60 (1.12–2.28) .009 NS NS 0.72 (0.63– 0.84) ⬍.001
Abbreviation: NS, not significant.
1.03 (1.01–1.06)
P Value
HR (95% CI)
P Value
1.24 (0.94 –1.65) 1.05 (1.02–1.07) 1.06 (0.81–1.40) 0.98 (0.71–1.36) 1.06 (0.71–1.59) 0.90 (0.66 –1.23) 0.80 (0.54 –1.17) 1.01 (0.95–1.07) 0.96 (0.92–1.00) 0.71 (0.60 – 0.83)
NS ⬍.001 NS NS NS NS NS NS NS ⬍.001
Multivariate (Adjusted) Analysis HR (95% CI)
P Value
NS .003 NS NS NS NS 1.64 (1.06 –2.55) .027 NS NS 0.76 (0.63– 0.91) .004
1.05 (1.02–1.08)
doi: 10.1210/jc.2012-3849
of the patients; however, the latter only could have been demonstrated by comparing FT3 values before admission with those at admission. FT3 was significantly decreased in patients admitted with sepsis compared with that in those admitted for other causes. This finding confirms the results of previous studies that showed a correlation of low T3 concentrations with the severity of illness in elderly hospitalized patients (23). All the factors that inhibit 5⬘-monodeiodinase, an enzyme involved in the peripheral 5⬘ deiodination of T4 to T3, such as malnutrition, low caloric intake, drugs, and increased inflammatory cytokines seem to be the main factors responsible for the development of low T3 syndrome during the hospitalization for acute illness (32–34). Some, but not all, reports have suggested that decreased TSH and increased FT4 levels are associated with an increase in mortality in elderly individuals (12–15, 18), whereas mild subclinical hypothyroidism appears to be related to an increase in survival in this population (13, 15, 16). A recent retrospective study performed in hospitalized patients older than 60 years showed that low total T4 and high TSH levels were associated with a worse prognosis for morbidity and mortality during hospitalization and for 18 months after discharge (19). To the best of our knowledge, our survey is the first study that has prospectively evaluated the relationship between the TFT results at hospital admission for acute illness and mortality in elderly patients in the long term. We found that alterations in TFT results were associated with mortality, but, unlike Mingote et al (19), we found that patients who died had serum levels of TSH and FT3 significantly lower than those of patients who survived, with no differences in FT4 levels. Moreover, survival analysis showed a significantly lower median survival time for the lowest tertiles of FT3, FT4, and TSH. Such differences may be related to the type of hormones measured (FT4 and FT3 in our study), patient selection criteria, and interference with drugs affecting thyroid function. There is a well-known association between serum T3 concentrations and mortality in critically ill patients (21, 35–37). In fact, a low T3 level has also been considered to be a significant predictive marker for mortality in several pathological states, including heart disease (35), acute stroke (38), and the need for hemodialysis (39). Our study confirms this association between low FT3 levels and allcause and CV mortality, being the most important predictor of 7-year CV mortality in octogenarian patients, even more than age. The relatively large number of patients and the long time of follow-up are the main strengths of this study. Our results related to CV mortality evaluation are limited by
jcem.endojournals.org
4689
the selection bias because we could not find the cause of death in a nonnegligible number of patients. In conclusion, our results show that alterations in TFT results are related to mortality in aged patients hospitalized for acute illness not only during hospitalization but also long term after discharge. Low FT3, low FT4, and low TSH serum concentrations are associated with decrease survival time, although only low FT3 behaves as a significant predictor of all-cause and CV mortality.
Acknowledgments Address all correspondence and requests for reprints to: Dr Pedro Iglesias, Department of Endocrinology, Hospital Ramón y Cajal, Ctra. de colmenar, Km 9,100, 28034 Madrid, Spain. E-mail: [email protected]. Disclosure Summary: The authors have nothing to disclose.
References 1. Chuang CC, Wang ST, Wang PW, Yu ML. Prevalence study of thyroid dysfunction in the elderly of Taiwan. Gerontology. 1998; 44:162–167. 2. Canaris GJ, Manowitz NR, Mayor G, Ridgway EC. The Colorado thyroid disease prevalence study. Arch Intern Med. 2000;160:526 – 534. 3. Hollowell JG, Staehling NW, Flanders WD, Hannon WH, Gunter EW, Spencer CA, Braverman LE. Serum TSH, T4, and thyroid antibodies in the united states population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab. 2002;87:489 – 499. 4. Valeix P, Dos Santos C, Castetbon K, Bertrais S, Cousty C, Hercberg S. Thyroid hormone levels and thyroid dysfunction of French adults participating in the SU.VI.MAX study. Ann Endocrinol (Paris). 2004;65:477– 486. 5. Empson M, Flood V, Ma G, Eastman CJ, Mitchell P. Prevalence of thyroid disease in an older Australian population. Intern Med J. 2007;37:448 – 455. 6. Brochmann H, Bjøro T, Gaarder PI, Hanson F, Frey HM. Prevalence of thyroid dysfunction in elderly subjects. A randomized study in a Norwegian rural community (Naeroy). Acta Endocrinol (Copenh). 1988;117:7–12. 7. Vanderpump MP, Tunbridge WM, French JM, et al. The incidence of thyroid disorders in the community: A twenty-year follow-up of the Whickham survey. Clin Endocrinol (Oxf). 1995;43:55– 68. 8. Aghini-Lombardi F, Antonangeli L, Martino E, et al. The spectrum of thyroid disorders in an iodine-deficient community: The Pescopagano survey. J Clin Endocrinol Metab. 1999;84:561–566. 9. Rehman SU, Cope DW, Senseney AD, Brzezinski W. Thyroid disorders in elderly patients. South Med J. 2005;98:543–549. 10. Cappola AR, Fried LP, Arnold AM, et al. Thyroid status, cardiovascular risk, and mortality in older adults. JAMA. 2006;295:1033– 1041. 11. Faggiano A, Del Prete M, Marciello F, Marotta V, Ramundo V, Colao A. Thyroid diseases in elderly. Minerva Endocrinol. 2011; 36:211–231. 12. Parle JV, Maisonneuve P, Sheppard MC, Boyle P, Franklyn JA. Prediction of all-cause and cardiovascular mortality in elderly people from one low serum thyrotropin result: a 10-year cohort study. Lancet. 2001;358:861– 865. 13. Gussekloo J, van Exel E, de Craen AJ, Meinders AE, Frölich M,
4690
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Iglesias et al
Thyroid and Mortality in the Elderly
Westendorp RG. Thyroid status, disability and cognitive function, and survival in old age. JAMA. 2004;292:2591–2599. Singer RB. Mortality in a complete 4-year follow up of 85-year-old residents of Leiden, classified by serum level of thyrotropin and thyroxine. J Insur Med. 2006;38:14 –19. Gussekloo J, van Exel E, de Craen AJ, Meinders AE, Frölich M, Westendorp RG. Thyroid function, activities of daily living and survival in extreme old age: the ’Leiden 85-plus Study.’ Ned Tijdschr Geneeskd. 2006;150:90 –96. van den Beld AW, Visser TJ, Feelders RA, Grobbee DE, Lamberts SW. Thyroid hormone concentrations, disease, physical function, and mortality in elderly men. J Clin Endocrinol Metab. 2005;90: 6403– 6409. Waring AC, Harrison S, Samuels MH, et al. Thyroid function and mortality in older men: A prospective study. J Clin Endocrinol Metab 2012;97:862– 870. Waring AC, Arnold AM, Newman AB, Bùzková P, Hirsch C, Cappola AR. Longitudinal changes in thyroid function in the oldest old and survival: the cardiovascular health study all-stars study. J Clin Endocrinol Metab. 2012;97:3944 –3950. Mingote E, Merono T, Rujelman R, et al. High TSH and low T4 as prognostic markers in older patients. Hormones (Athens). 2012;11: 350 –355. Bossoni S, Cossi S, Marengoni A, et al. Low T3 syndrome and outcome in elderly hospitalized geriatric patients. J Endocrinol Invest 2002;25(10 suppl):73–74. Iglesias P, Muñoz A, Prado F, et al. Alterations in thyroid function tests in aged hospitalized patients: prevalence, aetiology and clinical outcome. Clin Endocrinol (Oxf). 2009;70:961–967. Atkinson RL, Dahms WT, Fisher DA, Nichols AL. Occult thyroid disease in an elderly hospitalized population. J Gerontol. 1978;33: 372–376. Simons RJ, Simon JM, Demers LM, Santen RJ. Thyroid dysfunction in elderly hospitalized patients. Effect of age and severity of illness. Arch Intern Med. 1990;150:1249 –1253. Bossoni S, Cossi S, Marengoni A, et al. The negative role of subclinical thyrotoxicosis on the outcome of hospitalized geriatric patients. J Endocrinol Invest 2002;25(10 suppl):64 – 65. Díez JJ, Molina I, Ibars MT. Prevalence of thyroid dysfunction in adults over age 60 years from an urban community. Exp Clin Endocrinol Diabetes. 2003;111:480 – 485. Cobler JL, Williams ME, Greenland P. Thyrotoxicosis in institu-
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
27.
28.
29.
30.
31.
32.
33.
34. 35.
36.
37.
38.
39.
tionalized elderly patients with atrial fibrillation. Arch Intern Med. 1984;144:1758 –1760. Gencer B, Collet TH, Virgini V, et al. Subclinical thyroid dysfunction and the risk of heart failure events: an individual participant data analysis from 6 prospective cohorts. Circulation 2012;126: 1040 –1049. Sawin CT, Geller A, Wolf PA, et al. Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N Engl J Med. 1994;331:1249 –1252. Nanchen D, Gussekloo J, Westendorp RG, et al. Subclinical thyroid dysfunction and the risk of heart failure in older persons at high cardiovascular risk. J Clin Endocrinol Metab 2012;97:852– 861. Iida M, Yamamoto M, Ishiguro Y, Yamazaki M, Honjo H, Kamiya K. Thyroid hormone within the normal range is associated with left ventricular mass in patients with hypertension. J Am Soc Hypertens. 2012;6:261–269. Iglesias P, Muñoz A, Prado F, et al. Serum thyrotropin concentration is an early marker of normalization of low triiodothyronine syndrome in aged hospitalized patients after discharge. J Endocrinol Invest. 2010;33:607– 611. van der Poll T, Romijn JA, Wiersinga WM, Sauerwein HP. Tumor necrosis factor: a putative mediator of the sick euthyroid syndrome in man. J Clin Endocrinol Metab. 1990;71:1567–1572. Stouthard JM, van der Poll T, Endert E, et al. Effects of acute and chronic interleukin-6 administration on thyroid hormone metabolism in humans. J Clin Endocrinol Metab. 1994;79:1342–1346. Adler SM, Wartofsky L. The nonthyroidal illness syndrome. Endocrinol Metab Clin North Am 2007;36:657– 672, vi Iervasi G, Pingitore A, Landi P, et al. Low-T3 syndrome: a strong prognostic predictor of death in patients with heart disease. Circulation. 2003;107:708 –713. Peeters RP, Wouters PJ, van Toor H, Kaptein E, Visser TJ, Van den Berghe G. Serum 3,3⬘,5⬘-triiodothyronine (rT3) and 3,5,3⬘-triiodothyronine/rT3 are prognostic markers in critically ill patients and are associated with postmortem tissue deiodinase activities. J Clin Endocrinol Metab. 2005;90:4559 – 4565. Plikat K, Langgartner J, Buettner R, et al. Frequency and outcome of patients with nonthyroidal illness syndrome in a medical intensive care unit. Metabolism. 2007;56:239 –244. Alevizaki M, Synetou M, Xynos K, Pappa T, Vemmos KN. Low triiodothyronine: a strong predictor of outcome in acute stroke patients. Eur J Clin Invest. 2007;37:651– 657. Zoccali C, Mallamaci F. Thyroid function and clinical outcomes in kidney failure. Clin J Am Soc Nephrol. 2012;7:12–14.
ARTICLE
E n d o c r i n e
C a r e
Thyroid Function Tests and Mortality in Aged Hospitalized Patients: A 7-Year Prospective Observational Study Pedro Iglesias, Elena Ridruejo, Angélica Muñoz, Florentino Prado, María Cruz Macías, María Teresa Guerrero, Pilar Tajada, Carmen García-Arévalo, and Juan José Díez Department of Endocrinology (P.I., J.J.D.), Hospital Ramón y Cajal, 28034 Madrid, Spain; and Departments of Geriatrics (E.R., A.M., F.P., M.C.M., M.T.G.) and Biochemistry (P.T., C.G.-A.), Hospital General, 28007 Segovia, Spain
Context: Several alterations in thyroid function test (TFT) results have been associated with mortality in elderly patients. Objective: Our aim was to investigate the relationship between TFT results and all-cause and cardiovascular (CV) mortality in aged hospitalized patients. Design: A 7-year prospective observational study was conducted. TFTs were performed at hospital admission, and mortality was registered in the follow-up period. Patients: Participants were 404 patients aged ⬎65 years admitted to the Department of Geriatrics, Hospital General, Segovia, Spain, for any reason during 2005. Main Outcome Measures: The study evaluated the association between TFT results and mortality from all causes and CV diseases. Methods: TSH, free T4, and free T3 (FT3) were measured on the first day of admission. In-hospital and total survival times, number of deaths, and all-cause and CV mortality were registered until the census date on January 1, 2012. Results: During the study, 323 patients (80%) died. Kaplan-Meier analysis showed that median survival time for all-cause mortality was significantly lower in patients in the first tertile of serum FT3, in the first tertile of TSH, and in the first tertile of serum free T4 concentrations. Multivariate adjusted Cox regression analysis showed that the history of cancer (hazard ratio, 1.60; 95% confidence interval, 1.12–2.28; P ⫽ .009), age (1.03; 1.01–1.06; P ⫽ .003), and FT3 levels (0.72; 0.63– 0.84; P ⬍ .001) were significant factors related to all-cause mortality. The cause of death was known in 202 patients. Of this group, 61 patients (30.2%) died of CV disease. Patients in the first tertile of TSH and FT3 exhibited a significant higher mortality due to CV disease. In the adjusted Cox regression analysis, FT3 was a significant predictor of CV mortality (0.76; 0.63– 0.91; P ⫽ .004). Conclusions: Alterations in TFT results during hospitalization are associated with long-term mortality in elderly patients. In particular, low FT3 levels are significantly related to all-cause and CV mortality. (J Clin Endocrinol Metab 98: 4683– 4690, 2013)
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2013 by The Endocrine Society Received November 7, 2012. Accepted September 25, 2013. First Published Online October 30, 2013
doi: 10.1210/jc.2012-3849
Abbreviations: CHF, congestive heart failure; CHD, coronary heart disease; CV, cardiovascular; FT3, free T3; FT4, free T4; NTIS, nonthyroidal illness syndrome; TFT, thyroid function test.
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
jcem.endojournals.org
4683
4684
Iglesias et al
Thyroid and Mortality in the Elderly
hyroid dysfunction is a common disorder in the general population, mainly in women of advanced age (1–5). The prevalence of overt hypothyroidism and hyperthyroidism ranges between 1% and 7% and between 0.5% and 3.0%, respectively (2, 3, 6 – 8), increasing to 14% to 18% and higher than 5% when subclinical hypothyroidism and subclinical hyperthyroidism are considered (3, 8). Untreated thyroid disorders in elderly individuals are associated with significant morbidity (9 –11). Several studies (12–16), but not all (10, 17), suggest a relationship between altered thyroid function test (TFT) results and mortality in elderly individuals. Low levels of TSH or elevated levels of free T4 (FT4) have been related to an increased mortality rate in this population (12–15, 18). On the contrary, a high TSH level or low serum FT4 level was associated with lower mortality (13, 15, 16). A recent survey showed that a single measurement of thyroid function (TSH and T4) was not able to predict total or cause-specific mortality in a cohort of community-dwelling older men over an 8-year period of follow-up (17). Alterations in TFT results have also been associated with morbidity and mortality in older hospitalized patients. A recent retrospective study has shown that low T4 and high TSH levels were associated with a worse prognosis in these patients (19). It was also reported that a low serum T3 concentration behaves as a powerful predictor
T
Table 1.
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
for mortality during hospitalization in elderly patients (20, 21). In summary, some epidemiological studies suggest that different derangements in TFT results are related to mortality in elderly individuals in diverse clinical situations. Our aim has been to analyze whether a single TFT (TSH, FT4, and free T3 [FT3]) measurement could be a predictive factor of a better or worse vital prognosis not only during hospitalization but also in the long-term, as well as its relationship with all-cause and cardiovascular (CV) mortality in elderly patients admitted because of an acute process.
Patients and Methods Study design A 7-year (between January 1, 2005, and January 1, 2012) prospective observational study including patients older than 65 years hospitalized because of an acute process during 2005 was carried out. Patients were recruited from a previous cross-sectional study (21). All patients admitted to our geriatric unit were included in the study regardless of the cause of hospitalization. All patients taking medications known to modify thyroid function (antithyroid drugs, thyroxine, glucocorticoids, octreotide, dopamine, lithium, or amiodarone) were excluded. Demographic, clinical, and laboratory data were recorded (Table 1). In-hospital and total survival times, number of deaths,
Clinical and Analytical Data and Prevalent Comorbidities of the Study Patients
No. of patients, % Age, y Hypertension, % Diabetes, % Dyslipidemia, % Cardiovascular disease, % CHD Stroke Peripheral arterial disease CHF Arrhythmia Cancer, % Alterations in TFT results, % NTIS Subclinical hypothyroidism Overt hypothyroidism Subclinical hyperthyroidism Overt hyperthyroidism Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Thyrotropin, mU/L FT4, pmol/L FT3, pmol/L
Women
Men
Total
247 (61.1) 86.7 ⫾ 6.3 153 (61.9)a 67 (27.1) 37 (15.0) 84 (34.0) 18 (7.3) 46 (18.6) 8 (3.2) 45 (18.2) 94 (38.0) 26 (10.5) 180 (72.9) 151 (61.1) 16 (6.5) 2 (0.8) 5 (2.0) 6 (2.4) 27.4 ⫾ 6.2 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.6) 17.2 (15.2–19.7)a 3.7 ⫾ 0.9
157 (38.9) 84.8 ⫾ 6.3 67 (42.7) 34 (21.7) 23 (14.6) 46 (29.3) 20 (12.7) 19 (12.1) 18 (11.5) 16 (10.2) 52 (33.1) 28 (17.8) 115 (73.2) 100 (63.7) 7 (4.4) 2 (1.3) 3 (1.9) 3 (1.9) 27.2 ⫾ 3.5 130 (110 –140) 70 (60 – 80) 1.5 (0.8 –2.5) 16.3 (14.3–18.4) 3.5 ⫾ 0.9
404 85.9 ⫾ 6.4 220 (54.4) 101 (25.0) 60 (14.8) 130 (32.2) 38 (9.4) 65 (16.1) 26 (6.4) 61 (15.1) 146 (36.1) 54 (13.4) 295 (73.0) 251 (62.1) 23 (5.7) 4 (1.0) 8 (2.0) 9 (2.2) 27.3 ⫾ 5.1 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.5) 16.9 (14.8 –19.1) 3.6 ⫾ 0.9
Data represent the number of patients (percentage) and/or the mean ⫾ SD for normally distributed data and median (interquartile range) for nonnormally distributed data. a
Level of significance for multiple comparisons was adjusted using the Bonferroni correction to ␣: P ⫽ .004
doi: 10.1210/jc.2012-3849
and all-cause and CV mortality were registered until the census date on January 1, 2012. The vital status of every patient was evaluated through telephone, hospital medical reports, civil registration, and contact with the primary care physicians of the patients. The cause of death was obtained through the clinical information of the hospital medical reports and death certificates of the patients. CV death was defined as death from stroke, coronary heart disease (CHD), congestive heart failure (CHF), arrhythmia, myocardial infarction, or sudden death. This study was approved by our local ethics committee, and informed consent was given by subjects or their relatives before participation in the study.
Thyroid functional status Thyroid function was assessed by measuring serum concentrations of TSH, FT4, and FT3. Patients were classified according to the type and severity of thyroid dysfunction in overt hyperthyroidism (low TSH and high FT4 and/or high FT3), subclinical hyperthyroidism (low TSH and normal FT4 and FT3), overt hypothyroidism (high TSH and low FT4), and subclinical hypothyroidism (high TSH levels with normal FT4). Nonthyroidal illness syndrome (NTIS) or low T3 syndrome was also assessed.
Patients From an initial group of 447 patients, we excluded 43 patients who were taking thyroid-altering medications. We finally selected a group of 404 patients (Table 1). Clinical analytical data for the study patients are summarized in Table 1.
Hormone assays TSH, FT4, and FT3 serum concentrations were measured in all patients. Fasting samples of venous blood were obtained from an antecubital vein between 8:30 and 9:00 AM for hormonal quantifications on the first day of admission. Serum TSH, FT4, and FT3 were measured by an electrochemiluminescence immunoassay by using an E170 analyzer (Roche Diagnostics, Mannheim, Germany). Maximal intra- and interassay coefficients of variation as indicated in the package inserts of the commercial kits were 1.35% and 1.16% for TSH, 2.04% and 3.89% for FT4, and 2.87% and 10.17% for FT3. The sensitivities of the TSH, FT4, and FT3 assays were 0.005 mU/L, 0.3 pmol/L, and 0.4 pmol/L, respectively. Reference values were as follows: TSH, 0.4 to 5.0 mU/L; FT4, 11 to 23 pmol/L; and FT3, 3.9 to 6.8 pmol/L.
Statistical analysis For quantitative variables, results are expressed as mean ⫾ SD for normally distributed data and as median (interquartile range) for nonnormally distributed data. The normality of the data distribution was tested using the Kolmogorov-Smirnov test. Categorical variables are described as percentages. For comparisons of means between 2 groups of subjects, the Student t test was used for normally distributed data, and the Mann-Whitney test was used for nonnormally distributed data. For the comparison of ⬎2 independent groups, one-way ANOVA and Kruskal-Wallis tests were used for normally and nonnormally distributed data, respectively. For categorical comparisons, the 2 test or the Fisher exact test was used. Bonferroni correction to the ␣ level was applied when tests with multiple comparisons were performed. Survival time was estimated by the Kaplan-Meier method; the log rank test was used to compare arms. Unadjusted and stepwise multivariate Cox regression models were used to assess
jcem.endojournals.org
4685
the effects of several quantitative (age, TSH, FT4, and FT3) and qualitative (sex, hypertension, diabetes mellitus, dyslipidemia, CV disease, and cancer) variables on the risk of death. Hazard ratios (HRs) and their 95% confidence intervals (CIs) for allcause and CV mortality were estimated. Patients were classified according to tertiles of TSH, FT4, and FT3 levels obtained from 3 approximately equal groups based on the data analyzed. This analysis implies a possible bias or lack of generalization of these tertiles in other patient groups. Differences were considered significant when P ⬍ .05.
Results TFT results and morbidity TFT results were abnormal in most of the patients (73%); NTIS was the most common alteration (251 patients [62.1%]), followed by subclinical hypothyroidism (23 patients [5.7%]), overt hyperthyroidism (9 patients [2.2%]), subclinical hyperthyroidism (8 patients [2.0%]), and overt hypothyroidism (4 patients [1.0%]) (Table 1). Prevalent comorbidities are shown in Table 1. The presence of diabetes, dyslipidemia, CV disease, and cancer did not result in significantly different serum concentrations of TSH, FT4, and T3. Only hypertensive patients showed significantly higher serum FT4 levels than nonhypertensive subjects. The main causes for hospital admission were CHF (77 patients [19.1%]), stroke (64 patients [15.8%]), respiratory tract infection (54 patients [13.4%]), acute digestive hemorrhage (25 patients [6.2%]), CHD (23 patients [5.7%]), exacerbation of chronic obstructive pulmonary disease (23 patients [5.7%]), cancer (14 patients [3.5%]), sepsis (11 patients [2.7%]), urinary tract infection (11 patients [2.7%]), anemia (7 patients [1.7%]), syncope (7 patients [1.7%]), acute gastroenteritis (7 patients [1.7%]), confusion syndrome (6 patients [1.5%]), bradycardia (6 patients [1.5%]), cellulitis (5 patients [1.2%]), pancreatitis (3 patients [0.7%]), and other conditions (61 patients [15.1%]). No significant differences in TSH and FT4 were found among these groups. Patients with sepsis had the lowest FT3 levels (2.7 ⫾ 0.8 pmol/L), which were significantly different from those found in patients with stroke (3.9 ⫾ 1.0 pmol/L, P ⫽ .002) and CHD (3.8 ⫾ 0.9 pmol/L, P ⫽ .001). TFT results and mortality During the study, 323 patients (80%) died (median survival time, 9 months; interquartile range, 1–31 months; 61 patients [15.1%] died during hospitalization). Distribution of the patients according to the reference values of thyroid parameters in relation to vital status and time of death is shown in Table 2. The percentage of patients with low T3 levels was significantly higher in the group of pa-
4686
Iglesias et al
Table 2. Status
Thyroid and Mortality in the Elderly
Distribution of Patients According to Reference Values of Thyroid Parameters in Relation to Their Vital
Dead in the Hospital (n ⴝ 61)
TSH FT4 FT3
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
Dead After Discharge (n ⴝ 262)
Survivors (n ⴝ 81)
Low
Normal
High
Low
Normal
High
Low
Normal
High
P Value
7 (11.5) 3 (4.9) 52 (85.2)
52 (85.2) 57 (93.4) 9 (14.8)
2 (3.3) 1 (1.6) 0
24 (9.2) 7 (2.7) 170 (64.9)
221 (84.4) 238 (90.8) 91 (34.7)
17 (6.5) 17 (6.5) 1 (0.4)
1 (1.2) 3 (3.7) 37 (45.7)
72 (88.9) 74 (91.4) 44 (54.3)
8 (9.9) 4 (4.9) 0
.007 NS ⬍.001
Abbreviation: NS, not significant. Data represent number of patients (%). Normal range: TSH, 0.4 to 5.0 mU/L; FT4, 11 to 23 pmol/L, and FT3, 3.9 to 6.8 pmol/L.
tients who died in the hospital compared with those who did after discharge. Only 9 of 144 patients (6.2%) with normal FT3 levels died during hospitalization, making that finding highly predictive of survival. Patients who died were older and showed lower values of FT3 than patients who survived (Table 3). The presence of abnormal TFT results was positively associated with mortality (P ⫽ .001). Among patients who died, both TSH (1.2 [0.7–2.2] vs 1.6 [0.9 –2.6] mU/L, P ⫽ .047] and FT3 (3.2 ⫾ 0.8 vs 3.7 ⫾ 0.9 pmol/L, P ⬍ .001) were lower in those who died during hospitalization than among those who survived to hospital discharge. No significant differences in FT4 were found. Of the 323 patients who died, we only could find the cause of death in 202 patients (62.5%). CV disease was the main cause of death in 61 of these patients (30.2%). We did not find any significant difference in clinical and analytical data and TFT results among patients who died of CV disease and those who died of other causes. No differences were found in TSH, FT3, and FT4 among patients Table 3. Study
who died of CV disease during hospitalization and those who died after discharge. Kaplan-Meier analysis showed a median survival time for all-cause mortality of 3.0 (95% CI, 1.1– 4.9), 13.0 (6.1–19.8), and 19.0 (12.1–13.3) months for patients belonging to the first (FT3 ⱕ3.18 pmol/L), second (3.18 ⬎ FT3 ⱕ 3.96 pmol/L), and third (FT3 ⬎3.96 pmol/L) tertiles of FT3, respectively (P ⬍ .001) (Figure 1). In regard to TSH, median survival times for all-cause mortality were 3.0 (95% CI, 0.6 –5.4), 17.0 (9.9 –24.1), and 12.0 (5.7– 18.3) months for patients belonging to the first (TSH ⱕ1.06 mU/L), second (1.06 ⬎ TSH ⱕ 2.09 mU/L), and third (TSH ⬎ 2.09 mU/L) tertiles of TSH, respectively (P ⫽ .006) (Figure 1). In the same way, median survival times for all-cause mortality were 6.0 (3.6 – 8.4), 19.0 (12.8 – 25.1), and 11.0 (6.6 –15.4) months for patients belonging to the first (FT4 ⱕ 15.4 pmol/L), second (15.4 ⬎ FT4 ⱕ 18.2 pmol/L), and third (FT4 ⬎ 18.2 pmol/L) tertiles of FT4, respectively (P ⫽ .004) (Figure 1). Similar findings for TSH (log rank test 6.4, P ⫽ .041) and FT3 (log rank test
Clinical and Analytical Data of the Study Patients at Entry According to Their Vital Status at the End of the
Patients, n (%) Age, y Hypertension, % Diabetes, % Dyslipidemia, % CV disease, % CHD Stroke Peripheral arterial disease CHF Cancer, % Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Thyrotropin, mU/L FT4, pmol/L FT3, pmol/L
Alive
Dead
81 (20) 84.0 ⫾ 6.8 48 (59.3) 15 (18.5) 11 (13.6) 27 (33.3) 5 (6.2) 16 (19.7) 2 (2.5) 10 (12.3) 8 (9.9) 27.3 ⫾ 6.0 135 (120 –150) 75 (70 – 80) 1.8 (0.9 –2.8) 17.2 (14.9 –19.3) 4.0 ⫾ 0.9
323 (80) 86.4 ⫾ 6.2 172 (53.2) 86 (26.6) 49 (15.1) 103 (31.9) 33 (10.2) 49 (15.2) 24 (7.4) 51 (15.8) 46 (14.2) 27.4 ⫾ 5.0 130 (110 –150) 70 (60 – 80) 1.5 (0.8 –2.4) 16.8 (14.8 –19.0) 3.5 ⫾ 0.9
P Value .002 NS NS NS NS
NS NS NS NS NS NS ⬍.001
Abbreviation: NS, not significant. Data represent the number of patients (%) or the mean ⫾ SD for normally distributed data and median (interquartile range) for nonnormally distributed data. a
Level of significance for multiple comparisons was adjusted using the Bonferroni correction to ␣: P ⫽ .004
doi: 10.1210/jc.2012-3849
jcem.endojournals.org
4687
Figure 1. Kaplan-Meier survival analysis for all-cause (left panels; n ⫽ 404) and CV mortality (right panels; n ⫽ 283) in elderly patients admitted for acute illness stratified according to tertiles of TSH, FT4, and FT3.
25.6, P ⬍ .001), were found when mortality due to CV disease was analyzed. Survival analysis in relation to tertiles of FT4 did not show significant differences in this group of patients (Figure 1).
In an unadjusted Cox regression model, only age (years) (HR, 1.03; 95% CI, 1.01–1.0; P ⫽ .001) and FT3 (picomoles per liter) (0.71; 0.62– 0.82; P ⬍ .001) were significantly related to all-cause mortality; whereas in the
4688
Iglesias et al
Thyroid and Mortality in the Elderly
multivariate (adjusted) analysis, a history of cancer (1.60; 1.12–2.28; P ⫽ .009), age (1.03; 1.01–1.06; P ⫽ .003), and FT3 levels (0.72; 0.63– 0.84; P ⬍ .001) were significant factors related to all-cause mortality. In regard to CV mortality, only age (1.05; 1.02–1.07; P ⬍ .001) and FT3 levels (0.71; 0.60 – 0.83; P ⬍ .001) were significant factors in the unadjusted analysis. In the multivariate (adjusted) analysis FT3 (0.76; 0.63– 0.91; P ⫽ .004) behaved as a significant factor associated with CV mortality along with cancer (1.64; 1.06 –2.55; P ⫽ .027) and age (years) (1.05; 1.02–1.08; P ⫽ .003) (Table 4).
Discussion Our results clearly show a significant relationship between TFT results and mortality in aged hospitalized patients not only during hospitalization but also long term after hospital discharge. Low serum levels of TSH, FT4, and, in particular, FT3 were associated with increased all-cause mortality. Low FT3 was a significant predictor of all-cause mortality both before and after hospital discharge. In addition, low TSH and low FT3 levels were associated with higher CV mortality. However, multivariate analysis showed that only FT3 and age were independent predictors for CV mortality in this population. Lastly, we could not find any significant relationship between thyroid status and comorbidity with the exception of increased FT4 levels in hypertensive patients and decreased FT3 levels in patients who were admitted for sepsis. Several studies have investigated the prevalence of thyroid dysfunction, as well as its relationship with morbidity and mortality in elderly populations (12, 13, 15–18, 20 – 25). Some of them were performed in an apparently
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
healthy elderly population (12, 13, 15–18), whereas others were performed in hospitalized (20 –24) or institutionalized (26) elderly patients. Our findings confirm that the prevalence of alterations in TFT results in elderly hospitalized patients is notably high, about three quarters of patients, mainly with NTIS, as reported previously (20, 21, 23). The association between alterations in TFT results and comorbidity has been well established (27). In elderly individuals, some studies found a relationship between subclinical thyroid hyperfunction and atrial fibrillation but not with other CV disorders (10, 28). Others relate subclinical hyperthyroidism (TSH ⬍0.1 mU/L) and subclinical hypothyroidism (TSH ⬎10 mUl) with an increase in the risk of incident CHF in elderly patients with known CV risk factors or previous CV disease (27, 29). Our study shows no significant relationship between TFT results and prevalent CV disease or known risk factors for CV disease, including diabetes mellitus and dyslipidemia. Only hypertensive patients showed higher FT4 levels, although within the normal range, with FT3 and TSH levels similar to those of nonhypertensive patients. This observation is in accordance with a recent report showing a positive correlation between FT4 and left ventricular mass in hypertensive euthyroid patients (30). Low T3 syndrome affects ⬃60% to 66% of elderly patients hospitalized for acute illness (19, 21, 23). This alteration is usually transient, serum T3 levels normalizing in about 30% a month after hospital discharge (31). Our survey did not find any association between the presence of low T3 syndrome and the prevalent comorbidity, suggesting that this alteration would be more closely related to the acute illness than to the type of chronic conditions
Table 4. Nonadjusted and Adjusted HRs (with 95% CIs) for the Development of All-Cause and CV Mortality in Patients Classified According to Several Clinical and Analytical (TFTs) Variables All-Cause Mortality Unadjusted Analysis
CV Mortality Multivariate (Adjusted) Analysis Unadjusted Analysis
Variable
HR (95% CI)
P Value
HR (95% CI)
Male sex Age, y Hypertension Diabetes mellitus Dyslipidemia CV disease Cancer TSH, mU/L FT4, pmol/L FT3, pmol/L
1.05 (0.84 –1.31) 1.03 (1.01–1.05) 1.19 (0.96 –1.50) 1.01 (0.78 –1.29) 1.14 (0.84 –1.55) 0.97 (0.76 –1.24) 0.75 (0.54 –1.03) 1.00 (0.97–1.04) 0.98 (0.95–1.01) 0.71 (0.62– 0.82)
NS .001 NS NS NS NS NS NS NS ⬍.001
NS .003 NS NS NS NS 1.60 (1.12–2.28) .009 NS NS 0.72 (0.63– 0.84) ⬍.001
Abbreviation: NS, not significant.
1.03 (1.01–1.06)
P Value
HR (95% CI)
P Value
1.24 (0.94 –1.65) 1.05 (1.02–1.07) 1.06 (0.81–1.40) 0.98 (0.71–1.36) 1.06 (0.71–1.59) 0.90 (0.66 –1.23) 0.80 (0.54 –1.17) 1.01 (0.95–1.07) 0.96 (0.92–1.00) 0.71 (0.60 – 0.83)
NS ⬍.001 NS NS NS NS NS NS NS ⬍.001
Multivariate (Adjusted) Analysis HR (95% CI)
P Value
NS .003 NS NS NS NS 1.64 (1.06 –2.55) .027 NS NS 0.76 (0.63– 0.91) .004
1.05 (1.02–1.08)
doi: 10.1210/jc.2012-3849
of the patients; however, the latter only could have been demonstrated by comparing FT3 values before admission with those at admission. FT3 was significantly decreased in patients admitted with sepsis compared with that in those admitted for other causes. This finding confirms the results of previous studies that showed a correlation of low T3 concentrations with the severity of illness in elderly hospitalized patients (23). All the factors that inhibit 5⬘-monodeiodinase, an enzyme involved in the peripheral 5⬘ deiodination of T4 to T3, such as malnutrition, low caloric intake, drugs, and increased inflammatory cytokines seem to be the main factors responsible for the development of low T3 syndrome during the hospitalization for acute illness (32–34). Some, but not all, reports have suggested that decreased TSH and increased FT4 levels are associated with an increase in mortality in elderly individuals (12–15, 18), whereas mild subclinical hypothyroidism appears to be related to an increase in survival in this population (13, 15, 16). A recent retrospective study performed in hospitalized patients older than 60 years showed that low total T4 and high TSH levels were associated with a worse prognosis for morbidity and mortality during hospitalization and for 18 months after discharge (19). To the best of our knowledge, our survey is the first study that has prospectively evaluated the relationship between the TFT results at hospital admission for acute illness and mortality in elderly patients in the long term. We found that alterations in TFT results were associated with mortality, but, unlike Mingote et al (19), we found that patients who died had serum levels of TSH and FT3 significantly lower than those of patients who survived, with no differences in FT4 levels. Moreover, survival analysis showed a significantly lower median survival time for the lowest tertiles of FT3, FT4, and TSH. Such differences may be related to the type of hormones measured (FT4 and FT3 in our study), patient selection criteria, and interference with drugs affecting thyroid function. There is a well-known association between serum T3 concentrations and mortality in critically ill patients (21, 35–37). In fact, a low T3 level has also been considered to be a significant predictive marker for mortality in several pathological states, including heart disease (35), acute stroke (38), and the need for hemodialysis (39). Our study confirms this association between low FT3 levels and allcause and CV mortality, being the most important predictor of 7-year CV mortality in octogenarian patients, even more than age. The relatively large number of patients and the long time of follow-up are the main strengths of this study. Our results related to CV mortality evaluation are limited by
jcem.endojournals.org
4689
the selection bias because we could not find the cause of death in a nonnegligible number of patients. In conclusion, our results show that alterations in TFT results are related to mortality in aged patients hospitalized for acute illness not only during hospitalization but also long term after discharge. Low FT3, low FT4, and low TSH serum concentrations are associated with decrease survival time, although only low FT3 behaves as a significant predictor of all-cause and CV mortality.
Acknowledgments Address all correspondence and requests for reprints to: Dr Pedro Iglesias, Department of Endocrinology, Hospital Ramón y Cajal, Ctra. de colmenar, Km 9,100, 28034 Madrid, Spain. E-mail: [email protected]. Disclosure Summary: The authors have nothing to disclose.
References 1. Chuang CC, Wang ST, Wang PW, Yu ML. Prevalence study of thyroid dysfunction in the elderly of Taiwan. Gerontology. 1998; 44:162–167. 2. Canaris GJ, Manowitz NR, Mayor G, Ridgway EC. The Colorado thyroid disease prevalence study. Arch Intern Med. 2000;160:526 – 534. 3. Hollowell JG, Staehling NW, Flanders WD, Hannon WH, Gunter EW, Spencer CA, Braverman LE. Serum TSH, T4, and thyroid antibodies in the united states population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab. 2002;87:489 – 499. 4. Valeix P, Dos Santos C, Castetbon K, Bertrais S, Cousty C, Hercberg S. Thyroid hormone levels and thyroid dysfunction of French adults participating in the SU.VI.MAX study. Ann Endocrinol (Paris). 2004;65:477– 486. 5. Empson M, Flood V, Ma G, Eastman CJ, Mitchell P. Prevalence of thyroid disease in an older Australian population. Intern Med J. 2007;37:448 – 455. 6. Brochmann H, Bjøro T, Gaarder PI, Hanson F, Frey HM. Prevalence of thyroid dysfunction in elderly subjects. A randomized study in a Norwegian rural community (Naeroy). Acta Endocrinol (Copenh). 1988;117:7–12. 7. Vanderpump MP, Tunbridge WM, French JM, et al. The incidence of thyroid disorders in the community: A twenty-year follow-up of the Whickham survey. Clin Endocrinol (Oxf). 1995;43:55– 68. 8. Aghini-Lombardi F, Antonangeli L, Martino E, et al. The spectrum of thyroid disorders in an iodine-deficient community: The Pescopagano survey. J Clin Endocrinol Metab. 1999;84:561–566. 9. Rehman SU, Cope DW, Senseney AD, Brzezinski W. Thyroid disorders in elderly patients. South Med J. 2005;98:543–549. 10. Cappola AR, Fried LP, Arnold AM, et al. Thyroid status, cardiovascular risk, and mortality in older adults. JAMA. 2006;295:1033– 1041. 11. Faggiano A, Del Prete M, Marciello F, Marotta V, Ramundo V, Colao A. Thyroid diseases in elderly. Minerva Endocrinol. 2011; 36:211–231. 12. Parle JV, Maisonneuve P, Sheppard MC, Boyle P, Franklyn JA. Prediction of all-cause and cardiovascular mortality in elderly people from one low serum thyrotropin result: a 10-year cohort study. Lancet. 2001;358:861– 865. 13. Gussekloo J, van Exel E, de Craen AJ, Meinders AE, Frölich M,
4690
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Iglesias et al
Thyroid and Mortality in the Elderly
Westendorp RG. Thyroid status, disability and cognitive function, and survival in old age. JAMA. 2004;292:2591–2599. Singer RB. Mortality in a complete 4-year follow up of 85-year-old residents of Leiden, classified by serum level of thyrotropin and thyroxine. J Insur Med. 2006;38:14 –19. Gussekloo J, van Exel E, de Craen AJ, Meinders AE, Frölich M, Westendorp RG. Thyroid function, activities of daily living and survival in extreme old age: the ’Leiden 85-plus Study.’ Ned Tijdschr Geneeskd. 2006;150:90 –96. van den Beld AW, Visser TJ, Feelders RA, Grobbee DE, Lamberts SW. Thyroid hormone concentrations, disease, physical function, and mortality in elderly men. J Clin Endocrinol Metab. 2005;90: 6403– 6409. Waring AC, Harrison S, Samuels MH, et al. Thyroid function and mortality in older men: A prospective study. J Clin Endocrinol Metab 2012;97:862– 870. Waring AC, Arnold AM, Newman AB, Bùzková P, Hirsch C, Cappola AR. Longitudinal changes in thyroid function in the oldest old and survival: the cardiovascular health study all-stars study. J Clin Endocrinol Metab. 2012;97:3944 –3950. Mingote E, Merono T, Rujelman R, et al. High TSH and low T4 as prognostic markers in older patients. Hormones (Athens). 2012;11: 350 –355. Bossoni S, Cossi S, Marengoni A, et al. Low T3 syndrome and outcome in elderly hospitalized geriatric patients. J Endocrinol Invest 2002;25(10 suppl):73–74. Iglesias P, Muñoz A, Prado F, et al. Alterations in thyroid function tests in aged hospitalized patients: prevalence, aetiology and clinical outcome. Clin Endocrinol (Oxf). 2009;70:961–967. Atkinson RL, Dahms WT, Fisher DA, Nichols AL. Occult thyroid disease in an elderly hospitalized population. J Gerontol. 1978;33: 372–376. Simons RJ, Simon JM, Demers LM, Santen RJ. Thyroid dysfunction in elderly hospitalized patients. Effect of age and severity of illness. Arch Intern Med. 1990;150:1249 –1253. Bossoni S, Cossi S, Marengoni A, et al. The negative role of subclinical thyrotoxicosis on the outcome of hospitalized geriatric patients. J Endocrinol Invest 2002;25(10 suppl):64 – 65. Díez JJ, Molina I, Ibars MT. Prevalence of thyroid dysfunction in adults over age 60 years from an urban community. Exp Clin Endocrinol Diabetes. 2003;111:480 – 485. Cobler JL, Williams ME, Greenland P. Thyrotoxicosis in institu-
J Clin Endocrinol Metab, December 2013, 98(12):4683– 4690
27.
28.
29.
30.
31.
32.
33.
34. 35.
36.
37.
38.
39.
tionalized elderly patients with atrial fibrillation. Arch Intern Med. 1984;144:1758 –1760. Gencer B, Collet TH, Virgini V, et al. Subclinical thyroid dysfunction and the risk of heart failure events: an individual participant data analysis from 6 prospective cohorts. Circulation 2012;126: 1040 –1049. Sawin CT, Geller A, Wolf PA, et al. Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N Engl J Med. 1994;331:1249 –1252. Nanchen D, Gussekloo J, Westendorp RG, et al. Subclinical thyroid dysfunction and the risk of heart failure in older persons at high cardiovascular risk. J Clin Endocrinol Metab 2012;97:852– 861. Iida M, Yamamoto M, Ishiguro Y, Yamazaki M, Honjo H, Kamiya K. Thyroid hormone within the normal range is associated with left ventricular mass in patients with hypertension. J Am Soc Hypertens. 2012;6:261–269. Iglesias P, Muñoz A, Prado F, et al. Serum thyrotropin concentration is an early marker of normalization of low triiodothyronine syndrome in aged hospitalized patients after discharge. J Endocrinol Invest. 2010;33:607– 611. van der Poll T, Romijn JA, Wiersinga WM, Sauerwein HP. Tumor necrosis factor: a putative mediator of the sick euthyroid syndrome in man. J Clin Endocrinol Metab. 1990;71:1567–1572. Stouthard JM, van der Poll T, Endert E, et al. Effects of acute and chronic interleukin-6 administration on thyroid hormone metabolism in humans. J Clin Endocrinol Metab. 1994;79:1342–1346. Adler SM, Wartofsky L. The nonthyroidal illness syndrome. Endocrinol Metab Clin North Am 2007;36:657– 672, vi Iervasi G, Pingitore A, Landi P, et al. Low-T3 syndrome: a strong prognostic predictor of death in patients with heart disease. Circulation. 2003;107:708 –713. Peeters RP, Wouters PJ, van Toor H, Kaptein E, Visser TJ, Van den Berghe G. Serum 3,3⬘,5⬘-triiodothyronine (rT3) and 3,5,3⬘-triiodothyronine/rT3 are prognostic markers in critically ill patients and are associated with postmortem tissue deiodinase activities. J Clin Endocrinol Metab. 2005;90:4559 – 4565. Plikat K, Langgartner J, Buettner R, et al. Frequency and outcome of patients with nonthyroidal illness syndrome in a medical intensive care unit. Metabolism. 2007;56:239 –244. Alevizaki M, Synetou M, Xynos K, Pappa T, Vemmos KN. Low triiodothyronine: a strong predictor of outcome in acute stroke patients. Eur J Clin Invest. 2007;37:651– 657. Zoccali C, Mallamaci F. Thyroid function and clinical outcomes in kidney failure. Clin J Am Soc Nephrol. 2012;7:12–14.
