This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Samuels, M. H. Articles by Nelson, J. C. Search for Related Content PubMed PubMed Citation Articles by Samuels, M. H. Articles by Nelson, J. C.

Variable Effects of Nonsteroidal Antiinflammatory Agents on Thyroid Test Results

Division of Endocrinology, Diabetes, and Clinical Nutrition (M.H.S., K.P.), Oregon Health and Science University, Portland, Oregon 97201; Quest Diagnostics Nichols Institute (D.A., J.C.N.), San Juan Capistrano, California 92690; and Departments of Medicine and Pathology (J.C.N.), Loma Linda University School of Medicine, Loma Linda, California 92354

Address all correspondence and requests for reprints to: Dr. M. H. Samuels, Oregon Health and Science University, Division of Endocrinology, Diabetes and Clinical Nutrition, 3181 SW Sam Jackson Park Road, Mail Code CR 107, Portland, Oregon 97201. E-mail: samuelsm{at}ohsu.edu.

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
To investigate the effects of nonsteroidal antiinflammatory drugs (NSAIDs) on thyroid tests, 25 healthy subjects underwent a single-dose study and/or a 1-wk study. In the single-dose study, subjects received a single dose of one of six NSAIDs (aspirin, salsalate, meclofenamate, ibuprofen, naproxen, or indomethacin) at 0800 h. Total and free thyroid hormones and TSH were analyzed 0, 1, 2, 3, 4, and 8 h later. In the 1-wk study, subjects received one of six NSAIDs for 7 d. Thyroid hormones and TSH were analyzed at 0800 h each day. Total T₄ and total T₃ were measured by RIA, free T₄ and free T₃ were measured by equilibrium dialysis, and TSH was measured by immunometric assay. There were no changes in any hormones after a single dose or 1 wk of ibuprofen, naproxen, or indomethacin. Single-dose aspirin or salsalate decreased, whereas meclofenamate increased, various total and free thyroid hormone measurements. One week of aspirin or salsalate decreased total T₄, free T₄ (salsalate only), total T₃, free T₃, and TSH. These data confirm that aspirin, salsalate, and meclofenamate affect total and free thyroid hormone measurements and identify three NSAIDs that did not change thyroid tests. TSH remained within the normal range during acute or 1-wk administration of all of the NSAIDs.

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
IN HEALTHY SUBJECTS, circulating T₄ and T₃ are tightly bound to serum proteins (1). Only 0.02% of T₄ and 0.3% of T₃ circulate in the free form, but this small, unbound fraction is responsible for the biological activity of circulating thyroid hormones. Various medications displace thyroid hormones from protein-binding sites, inducing transient increases in free hormone concentrations and suppression of TSH levels (2, 3). These drug-induced changes can lead to inappropriate diagnostic and therapeutic decisions if they are not correctly understood.

Certain nonsteroidal antiinflammatory drugs (NSAIDs) can displace thyroid hormones from protein-binding sites, leading to problematic thyroid hormone measurements (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21). These studies vary in design, the type of NSAID given, dose and duration of treatment, and patient population (healthy vs. ill). Most studies used salicylate or mefenamic acid derivatives, whereas few studies investigated other widely used NSAIDs. In addition, some studies were performed before the development of the assays for TSH, free T₄, and free T₃ that are currently available. For these reasons, we designed a study to investigate and compare the effects of commonly prescribed NSAIDs on thyroid hormone and TSH measurements in healthy subjects. Our hypothesis was that some NSAIDs in current use would have less effect on free thyroid hormone and TSH levels than salicylate, salsalate, and meclofenamate.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Experimental design

The study was approved by the Oregon Health and Science University Institutional Review Board. Twenty-five healthy subjects were recruited (11 women and 14 men) who were ages 18–45 yr, were within 20% of ideal body weight and weighing between 50 and 90 kg, had no acute or chronic illnesses, and were receiving no medications. Each subject underwent a screening visit, including a history and physical examination, and a blood sample was drawn for human chorionic gonadotrophin-ß (women of childbearing potential), hematocrit, and chemistry battery. Each subject also underwent a TRH stimulation test, consisting of a baseline blood sample for free T₄, free T₃, and TSH, followed by an iv injection of 250 ìg TRH and repeat blood samples for TSH at 20, 30, and 60 min. If the baseline screening laboratory results and thyroid hormone evaluation (free T₄ and TSH) were normal, the subject was invited to participate in the single-dose NSAID studies and the 1-wk NSAID studies. Each subject participated in one, more than one, or all of the studies, depending on subject availability and preference. If a subject received more than one NSAID, the studies were separated by at least 6 weeks.

Single-dose NSAID study. The subject reported to the Clinical Research Center (CRC) at 0800 h after an overnight fast. A blood sample was sent for human chorionic gonadotrophin-ß (women of childbearing potential), and the subject received a single oral dose of one of the following NSAIDs: aspirin, salsalate, meclofenamate, ibuprofen, naproxen, or indomethacin. The selection of NSAID was randomly determined. The doses of the NSAIDs were chosen to replicate typical single doses taken by patients for acute, limited painful conditions (for example, acute muscle strain). Blood samples were drawn through an indwelling iv at 0 h (pre-NSAID ingestion) and 1 h, 2 h, 3 h, 4 h, and 8 h after NSAID ingestion. The subject was allowed to eat starting 2 h after the NSAID administration. To provide a control for possible circadian variation in hormone levels, 11 subjects had blood drawn at the same times but with no NSAID administration (control group). No TRH test was performed during the single-dose NSAID study.

One-week NSAID study. One week after the above study, the subject was started on the same NSAID that was administered in the single-dose study. One week was chosen as the washout period based on literature that reports 1 wk is adequate time for any acute effects of the single NSAID dose on thyroid hormone levels to abate. The subject received maximum recommended therapeutic doses of the NSAID at the recommended intervals used in clinical practice for 7 d. The subject returned to the CRC each morning for the next 7 d at 0800 h in the fasting state (before taking that day’s NSAID) and had blood drawn for free and total T₄, free and total T₃, and TSH measurements. A pill count was performed each day to assess compliance. On the morning of the eighth day, the subject did not take his or her NSAID dose but underwent a TRH stimulation test. Seven days was chosen as the time period adequate for attaining a steady-state drug level for all of the NSAIDs listed and observing changes in the hypothalamic-pituitary-thyroid axis, while avoiding side effects due to long-term use of NSAIDs.

NSAID doses

The types and doses of NSAIDs used in this study are listed in Table 1. The types of NSAIDs represent the spectrum of NSAIDs in common clinical usage in outpatient settings (22, 23). Aspirin and salsalate are members of one class of NSAIDs; ibuprofen and naproxen are members of a second class. Indomethacin and meclofenamate are each members of different classes of NSAIDs than aspirin/salsalate or ibuprofen/naproxen. Thus, this study included six NSAIDs and four different classes of NSAIDs. Meclofenamate is used in clinical practice less frequently than the other NSAIDs but was included because there are data that demonstrate effects on thyroid hormone testing. Thus, meclofenamate, aspirin, and salsalate represent positive controls for the other commonly used agents. NSAID doses were chosen as maximum recommended therapeutic doses of approximately equal antiinflammatory potency. Aspirin, ibuprofen, and naproxen are all available over the counter. The doses were not adjusted for body weight because, in clinical practice, patients are given NSAIDs at fixed doses, regardless of patient size. However, subjects at the extremes of weight (<50 or >90 kg) were excluded to avoid extreme variation in drug levels or increased risks of side effects.

Thyroid hormone levels were measured in all samples as follows. Free T₄ was measured by direct equilibrium dialysis using a modification of the method of Nelson and Tomei (24). This free T₄ assay involves a 13-fold dilution of small serum molecules during analysis. Free T₃ was measured by equilibrium tracer (indirect) dialysis, using a modification of the method of Lee and Pileggi (25). This free T₃ assay involves a 20-fold dilution of small serum molecules during analysis. TSH was measured by third-generation chemiluminescent immunometric assay. Total T₄ and total T₃ were measured by RIA. All sera from a single individual were assayed in triplicate in a single assay run for each analyte. Individual values are the means of triplicate determinations. All assays were performed at Quest Diagnostics Nichols Institute (San Juan Capistrano, CA). Normal ranges for these assays are as follows: total T₄ (5.6–13.7 µg/dl), total T₃ (87–180 ng/dl), free T₄ (0.8–2.7 ng/dl), free T₃ (230–420 pg/dl), and TSH (0.4–4.2 mU/l).

Thyroid hormone levels during NSAID use were compared with baseline levels (time 0 at 0800 h during the 1-d study, and day 0 at 0800 h for the 1-wk study) using repeated-measures ANOVA with Bonferroni t tests, with each subject as his or her own control. TSH responses to TRH (peak minus baseline) during NSAID use were compared with each subject’s baseline TRH test by paired t tests. To control for possible circadian variation in thyroid hormone concentrations during the single-dose study, TSH and thyroid hormones were measured at the same time points (0, 1, 2, 3, 4, and 8 h) in 11 healthy volunteers, matched for age, gender distribution, and body mass index, who did not receive any NSAIDs. Untreated control subjects were not used for the 1-wk study because blood samples were drawn at 0800 h every day, and circadian changes were not an issue. All results are presented as means ± SEM.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
The following numbers of subjects completed each study without significant side effects: aspirin, n = 12; salsalate, n = 11; meclofenamate, n = 11; ibuprofen, n = 10; naproxen, n = 8; and indomethacin, n = 12. The following numbers of subjects dropped out of the 1-wk studies after 1–4 d due to gastrointestinal upset: aspirin, n = 2; salsalate, n = 2; meclofenamate, n = 1; and indomethacin, n = 1. Thyroid hormone levels from subjects who dropped out were not included in the analysis. Changes in individual thyroid hormone levels during the study are summarized in the following sections. To clarify data presentation, the figures show only significant changes in thyroid hormone levels; data are not included if changes were insignificant.

Single-dose study

Total T₄, single-dose NSAID administration (Fig. 1, top). After single-dose salsalate administration, total T₄ levels were decreased at 2, 3, and 4 h (nadir 27% decrease at 3 h, from 6.6 ± 0.4 to 4.8 ± 0.3 µg/dl). Nine of the 11 subjects had a total T₄ level below the normal range at least once in the 8 h after single-dose salsalate administration. After single-dose meclofenamate administration, total T₄ levels were increased at 2 h (12% increase, from 6.5 ± 0.2 to 7.3 ± 0.3 µg/dl), although no subject had a total T₄ level above the normal range. Aspirin, ibuprofen, naproxen, and indomethacin had no effects on total T₄ levels, and there were no changes in the control group over 8 h.

View larger version (20K): [in this window] [in a new window]   FIG. 1. Mean total T4 measurements during the single-dose (8 h) study (top panel) and the 1-wk study (bottom panel). Mean results are graphed using a unique symbol for each NSAID shown in the legend. Measurements that were significantly different from time 0 (P < 0.05) are indicated by a letter at that time point above or below the graph. Each letter refers to a specific NSAID, as indicated in the legend. Only salsalate and meclofenamate are shown in the top graph, and only aspirin and salsalate are shown in the bottom graph because the other NSAIDs did not cause any significant changes in total T4 during the study.

Free T₄, single-dose NSAID administration (Fig. 2, top).

View larger version (16K): [in this window] [in a new window]   FIG. 2. Mean free T4 measurements during the single-dose (8 h) study (top panel) and the 1-wk study (bottom panel). Mean results are graphed using a unique symbol for each NSAID shown in the legend. Measurements that were significantly different from time 0 (P < 0.05) are indicated by a letter at that time point below the graph. Each letter refers to a specific NSAID, as indicated in the legend. Only salsalate is shown because the other NSAIDs did not cause any significant changes in free T4 during the study.

Total T₃, single-dose NSAID administration (Fig. 3, top).

View larger version (22K): [in this window] [in a new window]   FIG. 3. Mean total T3 measurements during the single-dose (8 h) study (top panel) and the 1-wk study (bottom panel). Mean results are graphed using a unique symbol for each NSAID shown in the legend. Measurements that were significantly different from time 0 (P < 0.05) are indicated by a letter at that time point above or below the graph. Each letter refers to a specific NSAID, as indicated in the legend. Only aspirin, salsalate, and meclofenamate are shown in the top graph and only aspirin and salsalate are shown in the bottom graph because the other NSAIDs did not cause any significant changes in total T3 during the study.

Free T₃, single-dose NSAID administration (Fig. 4, top).

View larger version (22K): [in this window] [in a new window]   FIG. 4. Mean free T3 measurements during the single-dose (8 h) study (top panel) and the 1-wk study (bottom panel). Mean results are graphed using a unique symbol for each NSAID shown in the legend. Measurements that were significantly different from time 0 (P < 0.05) are indicated by a letter at that time point above or below the graph. Each letter refers to a specific NSAID, as indicated in the legend. Only aspirin, salsalate, and meclofenamate are shown in the top graph and only aspirin and salsalate are shown in the bottom graph because the other NSAIDs did not cause any significant changes in free T3 during the study.

TSH, single-dose NSAID administration (Fig. 5, top).

View larger version (21K): [in this window] [in a new window]   FIG. 5. Mean TSH measurements during the single-dose (8 h) study (top panel) and the 1-wk study (bottom panel). Mean results are graphed using a unique symbol for each NSAID shown in the legend. Mean TSH measurements declined significantly during the 8-h study at all time points in all groups but were no different than the untreated control group. Therefore, for the sake of clarity, significant differences from time 0 are not indicated in the top panel. Measurements that were significantly different from time 0 (P < 0.05) in the bottom panel are indicated by a letter at that time point above or below the graph. Each letter refers to a specific NSAID, as indicated in the legend. Only aspirin and salsalate are shown in the bottom graph because the other NSAIDs did not cause any significant changes in TSH during the study.

Total T₄, 1-wk NSAID administration (Fig. 1, bottom). During 1 wk of aspirin administration, total T₄ levels were decreased at d 3–7 (nadir 26% decrease at d 5, from 7.2 ± 0.5 to 5.3 ± 0.3 ìg/dl). Seven of the eight subjects had a total T₄ level below the normal range at least once during aspirin administration. During 1 wk of salsalate administration, total T₄ levels were decreased at d 1–7 (nadir 45% decrease at d 6, from 7.5 ± 0.3 to 4.1 ± 0.4 µg/dl). All eight of the subjects had a total T₄ level below the normal range at least once during salsalate administration. Meclofenamate, ibuprofen, naproxen, and indomethacin had no effects on total T₄ levels over 1 wk.

Free T₄, 1-wk NSAID administration (Fig. 2, bottom). During 1 wk of salsalate administration, free T₄ measurements were decreased at d 3–7 (nadir 39% decrease at d 6, from 1.20 ± 0.08 to 0.73 ± 0.06 ng/dl). Five of the eight subjects had a free T₄ level below the normal range at least once during salsalate administration. Aspirin, meclofenamate, ibuprofen, naproxen, and indomethacin had no effects on free T₄ measurements over 1 wk.

Total T₃, 1-wk NSAID administration (Fig. 3, bottom). During 1 wk of aspirin administration, total T₃ measurements were decreased at d 3–7 (nadir 30% decrease at d 3, from 136 ± 6 to 95 ± 8 ng/dl). Five of the eight subjects had a total T₃ level below the normal range at least once during aspirin administration. During salsalate administration, total T₃ measurements were decreased at d 1–7 (nadir 51% decrease at d 6, from 138 ± 7 to 67 ± 8 ng/dl). All eight subjects had a total T₃ level below the normal range at least once during salsalate administration. Meclofenamate, ibuprofen, naproxen, and indomethacin had no effects on total T₃ measurements over 1 wk.

Free T₃, 1-wk NSAID administration (Fig. 4, bottom). During 1 wk of aspirin administration, free T₃ measurements were decreased at d 3–5 (nadir 26% decrease at d 3, from 340 ± 13 to 252 ± 20 pg/dl). Two of the eight subjects had a free T₃ level below the normal range at least once during aspirin administration. During 1 wk of salsalate administration, free T₃ measurements were decreased at d 2–7 (nadir 50% decrease at d 6, from 353 ± 34 to 175 ± 24 pg/dl). All eight subjects had a free T₃ level below the normal range at least once during salsalate administration. Meclofenamate, ibuprofen, naproxen, and indomethacin had no effects on free T₃ levels over 1 wk.

TSH, 1-wk NSAID administration (Fig. 5, bottom). During 1 wk of aspirin administration, TSH levels were decreased at d 2 and 3 (nadir 36% decrease at d 3, from 2.40 ± 0.36 to 1.53 ± 0.32 mU/l). During 1 wk of salsalate administration, TSH levels were decreased on d 2 and 3 (nadir 49% decrease at d 3, from 2.27 ± 0.24 to 1.15 ± 0.25 mU/l). No TSH levels were below the normal range during either aspirin or salsalate administration. Meclofenamate, ibuprofen, naproxen, and indomethacin had no effects on TSH levels over 1 wk. TSH responses to TRH were not altered after 1 wk of any of the NSAIDs (data not shown).

The data obtained after administration of the three NSAIDs that led to significant changes in one or more thyroid hormones are summarized in Table 2. Aspirin and salsalate induced decreases in mean total and free thyroid hormone measurements in both the single-dose and the 1-wk study, with associated changes in TSH during the 1-wk study but not the single-dose study. Meclofenamate caused acute increases in mean total T₄, total T₃, and free T₃ levels at 2 h but did not affect any mean thyroid hormone levels at 24-h intervals during 1 wk of administration. Ibuprofen, naproxen, and indomethacin did not lead to significant changes in levels of any of the thyroid hormones during either the single-dose or the 1-wk study.

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

Among NSAIDs, the most extensive studies of this process have used aspirin, salsalate, and meclofenamic acid derivatives. Faber et al. (17) reported that a single dose of oral salicylate given to healthy volunteers increased free T₄ and decreased TSH within 30 min. Wang et al. (19) documented decreases in total T₄ and T₃ that were maximal at 4 h and decreased TSH levels that were maximal at 8 h. McConnell (21) reported decreased total T₄, total T₃, and TSH in healthy subjects within 24 h of starting salsalate 1500 mg twice daily, effects that persisted for 3 d. In other studies, total T₄ and/or total T₃ were still decreased compared with baseline in subjects given 1 wk of salsalate or aspirin, although TSH was unchanged (4, 5, 16). In subjects who receive salicylates for longer than 1 wk, there are case reports of patients with low total T₄, although free thyroid hormone and TSH are reported to be normal (12, 15, 16). In vitro, aspirin and salsalate inhibit T₄ binding to proteins and increase free hormone levels at therapeutic drug concentrations (4, 13, 14, 26, 27), confirming a direct effect on hormone binding.

After single-dose aspirin administration, we found in the present study that total and free T₃ (but not total and free T₄) decreased by 12 and 15%, respectively, at 3 h, which is consistent with reduced T₄ to T₃ conversion. After single-dose salsalate, both total T₄ and total T₃ decreased by 27 and 31%, respectively, at 2–4 h (Table 2 and Figs. 1 and 3). This agrees with some of the studies summarized earlier. Measurements of free T₄ and free T₃ also decreased by 13 and 28%, respectively, a change not previously reported. It is possible that we missed early increases in free thyroid hormones (as shown by Faber et al. at 30 min, using indirect ultrafiltration, which does not dilute binding inhibitors in vitro) (17). It is also possible that the dilution of NSAID concentrations during dialysis resulted in underestimates of serum free hormone concentrations. We also show that TSH decreased acutely in subjects given aspirin or salsalate, and ours is the first study to use an untreated control group for comparison. Because untreated controls had similar decrements in TSH during the single-dose study (between 0800 h and 1600 h), these decrements cannot be attributed to aspirin or salsalate administration. During the 1-wk study, sera were obtained at the same time each day, and TSH concentrations were decreased during aspirin and salsalate administration (Fig. 5 and Table 2). During the 1-wk study, total and free T₄ were reduced during salsalate administration. Total T₄, but not free T₄, was reduced during aspirin administration (Figs. 1 and 2, and Table 2). We assume that the TSH suppression (Fig. 5) contributed to lower free T₄ measurements and that reduced T₄ to T₃ conversion contributed to lower free T₃ values when free T₄ was not reduced and to the reduction of free T₃ when it was proportionately greater than the reduction of free T₄. Reduced T₄ binding to proteins would account for the reduced total T₄ concentrations when free T₄ was not reduced (Table 2).

Other published studies address the effects of meclofenamic acid derivatives on thyroid hormones. In vivo, mefenamic acid decreased total T₄ by 20% and TSH by 45% at 24 h of administration in one study (10). A second study reported decreased TSH by 66% and blunted TRH stimulation tests at 2–4 d of fenclofenac administration (8). Longer periods of treatment with meclofenamic acid derivatives (10 d to >3 months) led to decreased total T₄ and total T₃, with normal free T₄, free T₃, and TSH in most studies (7, 8, 9, 11). In vitro, meclofenamic acid and its derivatives are potent acute inhibitors of T₄ and T₃ binding to thyroid binding proteins, leading to increased free hormone fractions in the equilibrium dialysis measurements (13, 14). In addition, one study showed that meclofenamate or fenclofenac directly suppressed TRH-stimulated TSH release from cultured rat anterior pituitary cells, suggesting a direct pituitary effect on TSH secretion (28).

Our study is the first to show an acute increase in free T₃ in vivo after a single dose of meclofenamate (37% at 3 h). However, we did not see similar changes in free T₄ over this time course, and we documented increased, rather than decreased, total T₄ and total T₃ during the single-dose meclofenamate study. TSH was no different than the circadian control during the 24-h study. None of the thyroid hormone measurements were altered during the 1-wk study of meclofenamate.

Fewer published studies address the short-term effects of other NSAIDs on thyroid hormone levels. Cremoncini et al. (29) reported that 12 d of flurbiprofen administration did not change total T₄ or total T₃ in healthy subjects. Ramey et al. (5) and Croxson et al. (30) reported that 2 d or 1 wk of indomethacin did not alter total T₄, total T₃, or TSH in healthy subjects. Finally, patients receiving various NSAIDs for longer than 3 wk have had normal thyroid hormones in published studies (11, 18, 20). Our results extend these reports for the commonly prescribed NSAIDS ibuprofen, naproxen, and indomethacin. We found that total or free thyroid hormone, TSH, and TRH responses were not altered after acute or 1-wk administration of high therapeutic doses of these three NSAIDs. This indicates that these three NSAIDs were not disruptive of normal thyroid hormone homeostasis and is consistent with the hypothesis that these agents are weaker inhibitors of thyroid hormone binding.

Although we documented significant changes in various thyroid hormone measurements during administration of some of the NSAIDs, a more relevant clinical question may be how often thyroid hormone measurements fall outside the laboratory normal range in patients receiving these agents. During short-term administration of aspirin or salsalate, total thyroid hormone measurements below the normal range were so common as to be expected in the majority of healthy subjects. Aspirin and salsalate also affect the accuracy of free thyroid hormone measurements, but fewer free hormone values were below the normal range. Overtly low total or free thyroid hormone measurements do not commonly occur with the other NSAIDs studied here. Increases in mean total T₄ and in mean total and free T₃ measurements occurred during the first 3 h after meclofenamate administration. However, most individual values remained within the normal range, with the exception of occasionally elevated free T₃ values. The explanation for these increased values is unknown. TSH measurements do not fall outside the normal range during short-term administration of any of the NSAIDs prescribed in this study.

There are several limitations to this study. One is the issue of NSAID effects on clinical laboratory free thyroid hormone measurements. All clinical laboratory free hormone assays dilute serum samples, in vitro, during free hormone determinations. This dilution reduces NSAID concentrations during analysis. When a NSAID has been inhibiting T₄ or T₃ binding to serum proteins in vivo, the reduction in concentration in vitro will diminish binding inhibition, thereby increasing hormone binding and reducing free hormone concentrations compared with in vivo levels. This results in underestimates of in vivo serum free hormone concentrations. Thus, the clinical laboratory free thyroid hormone measurements reported in this and other studies of NSAID effects may well be underestimates of endogenous free T₄ and T₃ concentrations, especially during peak drug concentrations. Only after these drugs have been cleared from the circulation will these measurements accurately reflect in vivo free hormone concentrations. This should be considered in interpreting reported data. (This does not occur with ultrafiltration free hormone assays that do not dilute serum samples during analysis).

Regarding other issues, we limited the duration of NSAID exposure to 1 wk because we did not believe the risks of more chronic NSAID ingestion were justified in healthy subjects. We did not recruit chronically ill subjects receiving NSAIDs for longer periods of time to avoid effects of a heterogeneous patient population, chronic illness, and other medications on thyroid hormone levels. Thus, we cannot state with certainty that more chronic NSAID use would lead to the same results. However, the relative lack of effect on serum TSH measurements in our study, combined with published reports in patients on long-term NSAIDs, gives us confidence that TSH, at least, remains a valid indicator of patients’ underlying thyroid status while receiving NSAIDS over the long term. We also used a fixed, high therapeutic dose of each NSAID. We did not perform dose-response studies due to resource limitations, concentrating instead on effects of maximal doses in common clinical use. It is unlikely that lower doses of the NSAIDs would lead to more positive results, but it is entirely likely that lower doses would lead to lesser effects. Finally, because we did not study ill subjects, it is possible that more abnormal thyroid hormone measurements occur in subjects who have superimposed nonthyroidal illness and NSAID use.

In summary, we extend earlier studies that show significant effects of aspirin, salsalate, and meclofenamate on total or free thyroid hormone measurements. We also report the first detailed data on the absence of measurable effects on thyroid hormone measurements attributable to ibuprofen, naproxen, and indomethacin at conventional doses. In all cases, TSH remains the optimal screening test for thyroid disease in subjects receiving NSAIDs because its measurements were less affected by these agents.

	Acknowledgments

 
We thank the Oregon Health and Science University General Clinical Research Center staff for their excellent patient care and sample collection.

	Footnotes

 
This work was supported by Grant NIH R29 DK 48366 and the Oregon Health and Science University General Clinical Research Center (Grant NIH GCRC M01-RR00334).

Abbreviation: NSAID, Nonsteroidal antiinflammatory drug.

Received November 30, 2002.

Accepted September 8, 2003.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Schussler GC 2000 The thyroxine-binding proteins. Thyroid 10:141–147[Medline]
2. Surks MI, Sievert R 1995 Drugs and thyroid function. New Engl J Med 333:1688–1694[Free Full Text]
3. Davies PH, Franklyn JA 1991 The effects of drugs on tests of thyroid function. Eur J Clin Pharmacol 40:439–451[CrossRef][Medline]
4. Larsen PR 1972 Salicylate-induced increases in free triiodothyronine in human serum. J Clin Invest 51:1125–1134
5. Ramey JN, Burrow GN, Spaulding SW, Donabedian RK, Speroff L, Frantz AG 1976 The effect of aspirin and indomethacin on the TRH response in man. J Clin Endocrinol Metab 43:107–114[Abstract/Free Full Text]
6. Dussault JH, Turcotte R, Guyda H 1976 The effect of acetylsalicylic acid on TSH and PRL secretion after TRH stimulation in the human. J Clin Endocrinol Metab 43:232–235[Abstract/Free Full Text]
7. Ratcliff WA, Hazelton RA, Thomson JA, Ratcliff GR 1980 The effect of fenclofenac on thyroid function tests in vivo and in vitro. Clin Endocrinol 13:569–575[Medline]
8. Kurtz AB, Capper SJ, Clifford J, Humphrey MJ, Lukinac L 1981 The effect of fenclofenac on thyroid function. Clin Endocrinol 15:117–124[Medline]
9. Capper SJ, Humphrey MJ, Kurtz AB 1981 Inhibition of thyroxine binding to serum proteins by fenclofenac and related compounds. Clin Chim Acta 112:77–83[CrossRef][Medline]
10. Koiqumi Y, Sato A, Yamada T, Inada M 1984 Effect of mefenamic acid on plasma protein-thyroid hormone interaction, monodeiodination of thyroxine, urinary excretion of tri-iodothyronine and thyrotropin regulation. Clin Exper Pharmacol Physiol 11:291–299[Medline]
11. Evans PJ, Woodhead JS, Weeks I, Scanlon MF 1987 Circulating TSH levels measured with an immunochemiluminometric assay in patients taking drugs interfering with biochemical thyroid status. Clin Endocrinol 26:717–721[Medline]
12. Kabadi UM, Danielson S 1987 Misleading thyroid function tests and several homeostatic abnormalities induced by "disalcid" therapy. J Am Geriatr Soc 35:255–257[Medline]
13. Lim CF, Bai Y, Topliss DJ, Barlow JW, Stockigt JR 1988 Drug and fatty acid effects on serum thyroid hormone binding. J Clin Endocrinol Metab 67:682–687[Abstract/Free Full Text]
14. Munro SL, Lim CF, Hall JG, Barlow JW, Craik DJ, Topliss DJ, Stockigt JR 1989 Drug competition for thyroxine binding to transthyretin (prealbumin): comparison with effects on thyroxine-binding globulin. J Clin Endocrinol Metab 68:1141–1147[Abstract/Free Full Text]
15. McConnell RJ 1992 Abnormal thyroid function test results in patients taking salsalate. JAMA 267:1242–1243[Abstract/Free Full Text]
16. Latham BB, Horst E, Nankin H, Lin T, Osterman J 1993 Case report: abnormal thyroid function tests in a patient and two normal volunteers treated with salsalate. Am J Med Sci 305:111–113[Medline]
17. Faber J, Waetjen I, Siersbaek-Nielsen K 1993 Free thyroxine measured in undiluted serum by dialysis and ultrafiltration: effects of non-thyroidal illness, and an acute load of salicylate or heparin. Clin Chim Acta 223:159–167[CrossRef][Medline]
18. Bishnoi A, Carlson HE, Gruber BL, Kaufman LD, Bock JL, Lidonnici K 1994 Effects of commonly prescribed anti-inflammatory drugs on thyroid hormone measurements. Am J Med 96:235–238[CrossRef][Medline]
19. Wang R, Nelson J, Wilcox RB 1998 Salsalate administration–a potential pharmacological model of the sick euthyroid syndrome. J Clin Endocrinol Metab 83:3095–3099[Abstract/Free Full Text]
20. Carlson HE, Kaell AT, Schulman, PE, Tan M, Bock JL 1999 Effects of several nonsteroidal antiinflammatory drugs on thyroid function tests. J Rheumatol 26:1855–1856[Medline]
21. McConnell RJ 1999 Changes in thyroid function tests during short-term salsalate use. Metabolism 48:501–503[CrossRef][Medline]
22. Brooks PM, Day RO 1991 Nonsteroidal antiinflammatory drugs–differences and similarities. N Engl J Med 324:1716–1725[Medline]
23. Greene JM, Winickoff RN 1992 Cost-conscious prescribing of nonsteroidal anti-inflammatory drugs for adults with arthritis. A review and suggestions. Arch Intern Med 152:1995–2002[Abstract/Free Full Text]
24. Nelson JC, Tomei RT 1988 Direct determination of free thyroxin in undiluted serum by equilibrium dialysis/radioimmunoassay. Clin Chem 34:1737–1744[Abstract/Free Full Text]
25. Lee ND, Pileggi VJ 1970 Measurement of "free" thyroxine in serum. Clin Chem 17:166–173
26. Wang R, Nelson JC, Wilcox RB 1999 Salsalate and salicylate binding to and their displacement of thyroxine from thyroxine-binding globulin, transthyretin, and albumin. Thyroid 9:359–364[Medline]
27. Goussis OS, Theodoropoulos TJ 1990 Dilantin and salicylate effects on hepatic thyroxine bio-availability and dialyzable thyroxine. Horm Metab Res 22:342–344[Medline]
28. Lim CF, Loidl NM, Kennedy JA, Topliss DJ, Stockigt JR 1995 Effect of loop diuretics and nonsteroidal antiinflammatory drugs on thyrotropin release by rat anterior pituitary cells in vitro. Metabolism 44:1008–1012[CrossRef][Medline]
29. Cremoncini M, Libroia A, Valente C, Cremoncini A, Della Croce F, Losa M, Lerro S, Franco PA, Longoni A 1984 Flurbiprofen and thyroid function tests. Br J Clin Pract 38:399–402[Medline]
30. Croxson MS, Hall TD, Jaramillo JE, Nicoloff JT 1977 Lack of indomethacin effect on thyroid function in man. J Clin Endocrinol Metab 44:748–751[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Samuels, M. H. Articles by Nelson, J. C. Search for Related Content PubMed PubMed Citation Articles by Samuels, M. H. Articles by Nelson, J. C.