Cross-Reactive Mechanism for the False Elevation of Free Triiodothyronine in the Patients Treated with Diclofenac

Endocrine Journal 2001, 48 (6), 717-722

NOTE

KEIZO KASONO, HAJIME HIKINO*, SHINJI FUJINO**, NoBUYUKI TAKEMOTO*, TosHIHIRO KAI*, KIYosHI YAMAGUCHI***, FUMio KONISHI* AND MASANOBU KAWAKAMI

Department of Endocrinology and Metabolism, Jichi Medical School, Omiya Medical Center, 1-847, Amanuma-cho, Saitama 330-8503, Japan *Department of Surgery , Jichi Medical School, Omiya Medical Center, 1-847, Amanuma-cho, Saitama 330-8503, Japan **Department of Pathology , Jichi Medical School, Omiya Medical Center, 1-847, Amanuma-cho, Saitama 330-8503, Japan ***Department of Neurology , Jichi Medical School, Omiya Medical Center, 1-847, Amanuma-cho, Saitama 330-8503, Japan

Abstract. We report three cases of patients exhibiting a false elevation of serum free triiodothyronine (FT3) as a result of a cross-reaction with diclofenac. The first case is a 66-yr-old woman with a long history of rheumatoid arthritis (RA). The patient was receiving diclofenac for the treatment of her RA. The patient was subsequently diagnosed as having thyroid papillary adenocarcinoma and received a subtotal thyroidectomy. After the operation, the patient exhibited postoperative hypothyroidism except for a gradual elevation of FT3. The other two patients also exhibited an elevated serum FT3 level after the administration of diclofenac. Serum FT3 levels in these patients decreased to normal or below normal after diclofenac administration was discontinued. In the first case, the elimination of immunoglobulin from the sera using polyethylene glycol precipitation did not reduce the FT3 level. In our hospital, Vitros ECi (enhanced chemiluminescence enzyme immunoassay) system and Vitros FT3 kit were used for FT3 assay. The patient's FT3 levels were normal or below normal when they were measured using other FT3 kits. FT3 was also detected when diclofenac was dissolved in a phosphate buffered saline. Therefore, we concluded that a cross-reaction between FT3 and diclofenac was the mechanism causing the false elevation of FT3 in these patients.

Key Words: Triiodothyronine, Thyroid, False elevation, Cross-reaction, Diclofenac (Endocrine Journal 48: 717-722, 2001)

CLINICAL tests for evaluating thyroid function in- tion tests [2-6]. Diclofenac is a commonly used clude the measurement of TSH, FT4, FT3, T4 and NSAID that binds to thyroid hormone binding pro- T3. Circulating thyroid hormones are tightly bound teins, leading to the release of FT3 or FT4. There- to several serum proteins, including thyroxine bind- fore, the drug slightly raises free thyroid hormone ing globulin (TBG), thyroxine binding prealbumin levels and reduces total T4 and T3 concentrations [7, (TBPA), and albumin [1]. Nonsteroidal anti- 8]. Autoantibodies against hormones being assayed, inflammatory drugs (NSAIDs) as well as various heterophilic antibodies to immunoglobulins used in other medications, are known to affect thyroid func- the assay kits and rheumatoid factors (RF) may also interfere with the accurate measurement of hormone levels [9-12]. Cross-reactions with administered Received: May 11, 2001 drugs [13] or serum proteins [14J are a common cause Accepted: August 15, 2001 of falsely elevated hormone concentrations. In this Correspondence to: Dr. Keizo KASONO, Department of report, we describe the false elevation of serum FT3 Endocrinology and Metabolism, Jichi Medical School, Omiya Medical Center, 1-847, Amanuma-cho, Saitama 330-8503, levels assayed by Vitros FT3 kit in three patients. Japan We concluded that a cross-reaction with diclofenac 718 KASONO et al. was the mechanism for the false elevation of FT3. In thyroglobulin antibodies and anti-microsomal anti- the first case, the false elevation of FT3 concentra- bodies were negative, TSH receptor antibody was tions in the patient's sera was affected by the pa- 1.2% (normal 0-15%), thyroglobulin was 6.6 pg/L tient's thyroid function. In severe hypothyroid (normal 0-30 ag/L), RF was 944 IU/ml, T-Chol was state, the elevation of FT3 was even more remarka- 4.09 mmol/L and ALP was 161 U/L. On the basis ble. An inverse relationship between FT3 level and of this data, we eliminated the diagnosis of autoim- thyroid function was thus observed in the first pa- mune thyroiditis and diagnosed her as postoperative tient. hypothyroidism with falsely evaluated FT3 concentration. Since the patient complained of slight general fatigue, we started to prescribe T4 (50 Case Report pg/day) for hypothyroidism, beginning on April 21, 1999. After starting medication, thyroid function Case 1 gradually returned to normal and the general fatigue A 66-year-old woman with a 30-year history of RA disappeared. On January 17, 2000, serum FT4 was was diagnosed as having thyroid papillary adenocar- 22.7 pmol/L, FT3 was 9.39 pmol/L and TSH was cinoma. A subtotal thyroidectomy was performed 4.02 mU/L. on January 27, 1999. Before operation, serum FT4 was 17.6 pmol/L (normal, 10.4-20.3 pmol/L), FT3 Interpretation of the mechanism for the false ele- was 8.45 pmol/L (normal, 4.27-8.12 pmol/L), and vation of FT3 in case 1 TSH was 1.48 mU/L (normal, 0.47-4.71 mU/L). RF was 608 IU/ml (normal, 0-15 IU/ml), total cho- To elucidate the mechanism of the false elevation lesterol (T-chol) was 3.44 mmol/L (normal, 3.29- in serum FT3, we performed several examinations. 5.96 mmol/L) and alkaline phosphatase (ALP) was The patient agreed orally to an additional blood 162 U/L (normal, 89-285 U/L). Although the FT3 sampling for the purpose of this study. level was slightly higher than normal, the TSH was In our hospital, thyroid function including FT3, not suppressed, and she had no signs of thyrotoxi- FT4, T3, T4 and TSH was routinely assayed using cosis. She had previously undergone an operation Vitros Eci (enhanced chemiluminescence enzyme im- for a carcinoma in her left breast in 1997. At the munoassay) system. These kits were purchased from time of thyroidectomy, she was taking the following Ortho-Clinical Diagnostics Inc. (Cardiff, UK). The medications: long-acting diclofenac (75 mg/day), principle of FT3 measurement in Vitros FT3 kit is tamoxifen (20 mg/day), retinol palmitate (30,000 competitive immune reaction using peroxidase- U/day), methotrexate (7.5 mg/week), active form of labelled sheep anti-FT3 polyclonal antibodies. When vitamin D (0.75 pg/day), Vitamin C (600 mg/day), the patient's stocked sera was assayed using the and tocopherol nicotinate (300 mg/day). After sub- Amerlex FT3 RIA kit from the same company or total thyroidectomy, serum FT4 concentration grad- using kits purchased from other companies, the FT3 ually decreased and TSH level increased remarkably concentrations were lower than normal (data not (Table 1). On April 20, 1999, FT4 was 9.27 pmol/L, shown). These results confirmed the diagnosis of FT3 was 12.1 pmol/L, TSH was 75.9 mU/L, anti- postoperative hypothyroidism in the patient.

Table 1. Thyroid function of case 1 FALSE ELEVATION OF FREE TRIIODOTHYRONINE 719

Polyethylene glycol (PEG) precipitation was em- new Vitros FT3 kit may solve the disturbance with ployed to remove the immunoglobulin fraction from diclofenac in the assay. The T3 and FT4 concentra- the patient's serum. We used stocked samples col- tions assayed by the Vitros kits were also slightly lected on three different days: January 22, March 19 elevated after the administration of diclofenac. As and April 20, 1999. The FT3 concentrations were previously reported, serum TSH was weakly sup- reassayed using the Vitros FT3 kit before the PEG pressed after diclofenac administration [8, 16]. precipitation and values of 5.84, 6.67 and 7.68 These results suggested that a cross-reaction be- pmol/L were obtained, respectively. A slight reduc- tween FT3 and diclofenac or its metabolites might tion in FT3 levels occurred in the samples that had have caused a false elevation of FT3 level. To been stocked for a long period of time. After the confirm this mechanism, cyclooxigenase-2 (Cox-2), a PEG precipitation, the FT3 values for these three target enzyme of diclofenac, was used to neutralize samples were 5.59, 6.93 and 8.45 pmol/L, respec- the drug. First, two normal volunteers were ad- tively. The levels before and after PEG precipitation ministered by diclofenac (75 mg/day) for 3 days. showed no significant difference. This showed that One-hundred microliters of their sera were added to RF and other immunoglobulins were not involved in frozen tubes containing 10 it of a Cox-2 reaction the mechanism that produced the false elevation of solution containing 300 iM of diethyldithiocarba- FT3 level. mate (Wako, Osaka, Japan), 1.0% trehalose (Wako) To evaluate the effect of diclofenac on thyroid and 1 U of Cox-2 (Alexis Biochemicals, San Diego, function in the patient, the drug was removed from CA). The mixture of serum and Cox-2 was im- the patient's prescription. Diclofenac was replaced mediately used for FT3 assay using the Vitros FT3 with loxioprofen for seven days. The patient was kit. Before Cox-2 treatment, their FT3 concentra- then given one capsule of long-acting diclofenac (37.5 tions assayed by Vitros FT3 kit were 16.2 and 11.4 mg/C) in the morning, and serum samples were taken pmol/L, respectively. After the addition of Cox-2 to before and 1 to 8 hours after the administration. their sera, the FT3 values decreased to 6.98 and 6.34 Table 2 shows the high concentration of FT3 that was pmol/L, respectively. detected by the Vitros FT3 kit after the admin- Finally, we measured the FT3 level in a diclofenac istration of the long-acting diclofenac. The highest solution. Diclofenac powder was dissolved in di- FT3 value assayed by the kit, 12.0 pmol/L, was very methyl sulfoxide (DMSO) and the described con- similar to the value obtained in the outpatient clinic centrations were adjusted by dilution with a phos- on April 20. The patient's serum samples were phate-buffered saline (PBS) pH 7.4, containing 0.1% usually obtained in the morning, about 2-3 hours bovine serum albumin. The final concentration of after she had taken her prescribed drugs. Other kits, DMSO was 1.00-0.03%. The samples were assayed including a new version of the Vitros FT3 kit (new using the Vitros FT3 kit and the New Vitros FT3 kit. Vitros FT3 kit), were not affected by the admin- Diclofenac increased the FT3 concentration meas- istration of the drug as previously reported [15]. The ured by the Vitros FT3 kit in a dose-dependent different anti-FT3 polyclonal antibodies employed in manner (Table 3). The new Vitros FT3 kit did not

Table 2. Thyroid function after the administration of diclofenac 720 KASONO et al.

Table 3. Measurement of FT3 in the diclofenac solution fenac (50 mg/day) was prescribed. Four days after the administration of the drug, her FT4 was 11.7 pmol/L and FT3 increased to 10.8 pmol/L. Her serum FT3 level decreased to 4.27 pmol/L after the withdrawal of diclofenac from her medication.

Discussion

Commonly prescribed drugs, including nonsteroidal anti-inflammatory drugs (NSAIDs), anti-convul- sants, beta-adrenoreceptor antagonists, steroid hormones and heparin may induce abnormal thyroid produce a false elevation in FT3 at lower concentra- function tests in the absence of any clinical signs tions of diclofenac. Vitros FT3 kit detected some of thyroid dysfunction [8, 16]. FT3 (1.31 pmol/1) even in the diclofenac-free sample, In the first patient, the high FT3 concentrations in but the new kit did not detect any FT3 in the same a patient with postoperative hypothyroidism were sample. remarkable. Diclofenac, as well as salicylates or Although the patient continued to receive the same fenclofenac, prescribed for RA usually affect thyroid dosage of diclofenac, the FT3 level gradually de- function by releasing FT3 or FT4 from TBG or other creased after the administration of T4 was initiated. thyroid hormone binding proteins. Consequently, FT3 concentrations on June 18, 1999 and January 17, FT3 and (or) FT4 levels slightly increase while TSH 2000 were 10.2 and 9.39 pmol/L, respectively. levels decrease. Unlike this common pattern of discrepancy in patients treated with NSAIDs, this case Case 2 exhibited a typical postoperative hypothyroid state except for a high level of FT3. A 78-year-old woman who complained of muscle When FT3 concentrations were examined before weakness was admitted to our hospital. She was and after diclofenac administration in the first case, diagnosed as Guillain-Barre syndrome. Before her only the Vitros FT3 kit produced a false elevation in diagnosis, thyroid function was assayed to determine FT3 levels. FT3 concentrations were correlated with the cause of muscle weakness. Before the thyroid the time course for the serum concentration of function assay, the patient had received diclofenac diclofenac described in the manufacturer's com- (50 mg/day) for the treatment of lumbar pain caused ments. A cross-reaction between FT3 and diclofenac by osteoporosis. Thyroid function assay produced is the most likely reason for the false elevation in FT3 unexpected results: TSH was 2.06 mU/L, FT4 was detected by the Vitros FT3 kit. The slight elevation 12.6 pmol/L, and FT3 was 22.8 pmol/L. Despite in total T3 and FT4 after the administration of the remarkable elevation in serum FT3, she did not diclofenac also supports our speculation of cross- show any symptoms of hyperthyroidism. After reaction between thyroid hormones and diclofenac. replacing diclofenac with loxioprofen, FT3 level de- T3, FT3 and FT4 levels measured by Amerlex kits creased to 8.24 pmol/L. were not affected by the administration of diclofenac. Direct measurements of diclofenac dissolved in Case 3 PBS using the Vitros FT3 kit showed a dose- dependent increase in FT3 concentration (Table 3). A 54-year-old woman with chronic thyroiditis for Maximal concentration of diclofenac in the serum several years underwent a follow-up examination in after the administration of a 25 mg tablet of the drug our outpatient clinic. Her thyroid function was was usually 0.1-0.9 pg/ml [17]. The concentration normalized by the administration of 50 [cg/day of of the drug in PBS was very high compared to the T4. Her FT4 was 15.2 pmol/L and FT3 was 4.48 physiological concentration in the patient's serum. pmol/L. She complained of headaches, and diclo- One possible cause of this discrepancy may be the low FALSE ELEVATION OF FREE TRIIODOTHYRONINE 721

solubility of the drug in DMSO. The kit also de- FT3 value was observed just prior to the prescription tected FT3 even in the diclofenac-free solution. The of T4. At that time, TSH was 70.7 mU/L and FT4 value, 1.31 pmol/L, was higher than the minimal was 9.27 pmol/L. After T4-treatment was started, detectable concentration (0.6 pmol/L) of the kit. the falsely elevated FT3 values gradually de- Therefore, there was some problem with this in vitro creased. An inverse relationship was observed be- assay by Vitros FT3 kit, the cause of which was not tween the false elevation of FT3 and thyroid function clear. However, the new Vitros FT3 kit did not de- as assessed by the FT4 or TSH levels of the patient. tect FT3 in the diclofenac-free medium. We reported three cases showing false elevation of The NSAID, fenoprofen, has been previously FT3 assayed by Vitros FT3 kit, and examined the reported to cause a false elevation in T3 and FT3 mechanism of the phenomenon. In all three cases, levels via a cross-reaction mechanism [13]. Our cross-reaction of anti-FT3 antibodies against diclo- results suggest that a cross-reaction is the main cause fenac was considered to be a mechanism. Different of the false elevation in FT3 levels assayed by Vitros anti-FT3 antibodies are used in the new version of FT3 kit in case 1. The thyroid function of the other Vitros FT3 kit, and this may be the reason the kit was two cases before and after diclofenac administration not disturbed by diclofenac treatment. In the near also supports this mechanism. In the first case, FT3 future, new Vitros FT3 kit will be commercially levels gradually increased after subtotal thyroidec- available in Japan. tomy. One possible explanation for this finding is that the patient's hypothyroid state induced a delay in drug metabolism and clearance. The patient Acknowledgments received the same dosage of long-acting diclofenac before and after subtotal thyroidectomy. FT3 value The authors are grateful to Ortho-Clinical Diag- before the operation was slightly higher than the up- nostics, Inc., for providing the thyroid function as- per normal level. Presumably, this value was over- say kits and for information on the new version of estimated by cross-reaction with diclofenac. Peak Vitros FT3 kit.

References

1. Robbins J (1991) Thyroid hormone transport proteins 1(8166): 487-488. and the physiology of hormone binding. In: Braver- 7. Cavalieri RR, Pitt-Rivers R (1981) The effects of drugs man LE, Utiger RD (eds) The Thyroid 6th edition, on the distribution and metabolism of thyroid hor- J. B. Lippincott Company, Philadelphia, 111-120. mones. Pharmacol Rev 33: 55-79. 2. Larsen PR (1972) Salicylate-induced increases in free 8. Bishnoi A, Carlson HE, Gruber BL, Kaufman LD, triiodothyronine in human serum. Evidence of inhi- Bock JL, Lidonnici K (1994) Effects of commonly bition of triiodothyronine binding to thyroxine-bind- prescribed nonsteroidal anti-inflammatory drugs on ing globulin and thyroxine-binding prealbumin. J thyroid hormone measurements. Am J Med 96: 235- Clin Invest 51: 1125-1134. 238. 3. Capper SJ, Humphrey MJ, Kurtz AB (1981) Inhibi- 9. Kohse KP, Wisser H (1990) Antibodies as a source of tion of thyroxine binding to serum proteins by fen- analytical errors. J Clin Chem Clin Biochem 28: 881- clofenac and related compounds. Clin Chim Acta 112: 892. 77-83. 10. Boscato LM, Stuart MC (1988) Heterophilic anti- 4. Ratcliffe WA, Hazelton RA, Thompson JA, Ratcliffe bodies: a problem for all immunoassays. Clin Chem JG (1980) The effect of fenclofenac on thyroid func- 34: 27-33. tion tests in vivo and in vitro. Clin Endocrinol (Oxf ) 11. Norden AGW, Jackson RA, Norden LE, Griffin AJ, 13: 569-575. Barnes MA, Little JA (1997) Misleading results from 5. Kurts AB, Capper SJ, Clifford J, Humphrey MJ, immunoassays of serum free thyroxine in the presence Lukinac L (1981) The effect of fenclofenac on thyroid of rheumatoid factor. Clin Chem 43: 957-962. function. Clin Endocrinol 15: 117-124. 12. Despres N, Grant AM (1998) Antibody interference in 6. Humphrey MJ, Capper SJ, Kurtz AB (1980) Fenclo- thyroid assays: a potential for clinical misinforma- fenac and thyroid hormone concentrations. Lancet tion. Clin Chem 44: 440-454. 722 KASONO et al.

13. Tillman J, Leask JT, Campbell J, Logue FC, Frasser Vitros FT3 kit not affected by nonsteroidal anti-inflam- WD (1986) Effect of fenoprofen on thyroid function matory drug, diclofenac. Radioisotopes 49: 49-50. tests. Clin Chim Acta 161: 233-238. 16. Davies PH, Franklyn JA (1991) The effects of drugs 14. Kasono K, Sato K, Suzuki T, Ohmura E, Demura R, on tests of thyroid function. Eur J Clin Pharmacol 40: Shizume K, Tsushima T, Demura H (1991) Falsely 439-451. elevated serum parathyroid hormone levels due to 17. Aoyama A, Natori Y, Yamaguti N, Koike S, Kusa- immunoglobulin G in a patient with idiopathic hypo- kabe K, Demura R, Demura H (1990) The effects of parathyroidism. J Clin Endocrinol Metab 72: 217-222. diclofenac sodium on thyroid function tests in vivo 15. Ito T, Takimoto T (2000) Study of the improved and in vitro. Jpn J Clin Pathol 38: 688-692.