Thyroid function tests: often justi ed in the acutely ill

Transcription

1 Original Article Ann Clin Biochem 2000; 37: 158±164 Thyroid function tests: often justi ed in the acutely ill E J Lamb 1 and J Martin 2 From the Departments of 1 Clinical Biochemistry and 2 Clinical Audit, East Kent Hospitals NHS Trust, Kent and Canterbury Hospital, Ethelbert Road, Canterbury CT1 3NG, UK SUMMARY. It is claimed that inappropriate requesting of thyroid function tests (TFTs) is common in acutely ill patients. Consecutive inpatient TFTs (n=129) were assessed in relation to clinical history and common symptoms and signs of thyroid disease. Requests were justi ed in 69% of cases, most commonly on the basis of atrial brillation and/or tachycardia. There were no clear reasons for requesting TFTs in the remaining cases, although the yield of abnormal results in these patients was similar to that in those with justi ed requests. Thyroid stimulating hormone (TSH) concentration was increased (median 7 5 mu/l, range 4 8±38 6 mu/l) in 22 patients, six of whom had biochemical and/or clinical evidence of (previously undiagnosed) and ve of whom had pre-existing. Of the remaining 11 patients with increased TSH levels, three were con rmed to have compensated ; non-thyroidal (NTI) (including the effect of drugs) accounted for four cases. In four patients (one of whom died during the admission) follow-up was not possible. Of six patients with reduced TSH concentration (range 0 05±0 35 mu/l), one was thyrotoxic on carbimazole, one was receiving thyroxine for, one had NTI and three were lost to follow-up (two of whom died during their admission). Manifestations of thyroid disease are protean and often subtle, and TFTs are thus clinically justi ed in many unwell inpatients. Although NTI contributes to some cases of abnormal TSH levels, a signi cant number of TFT abnormalities are consistent with underlying thyroid abnormality requiring investigation/treatment. Biochemical assessment of thyroid function in patients with non-thyroidal (NTI) is dif cult, given that may cause both increased and decreased concentrations of thyroid stimulating hormone (TSH, thyrotrophin) and decreased concentrations of thyroid hormone. 1 Additionally, many drugs are known to interfere with thyroid function tests (TFTs), through both in vivo and in vitro actions. 2 In one oft-quoted study of a hospitalized population, 3 patients with low TSH concentrations were more likely to have NTI or to be receiving glucocorticoids than to have thyroid disease and, among those with a TSH concentration in the range 6 8±20 mu/l, NTI was six times as common as intrinsic thyroid disease. In view of this, the consensus view in the literature 2,4±6 has been that TFTs should not be requested in ill patients unless there are clinical Correspondence: Dr E J Lamb. edmund.lamb@kch-tr.sthames.nhs.uk grounds for suspecting that thyroid dysfunction may be contributing to their clinical condition. Additionally, it is a widely held belief that unnecessary requesting of TFTs is common among patients admitted to hospital. 1,4,5 We have assessed the appropriateness of thyroid function testing in an inpatient population, in the context of common symptoms and signs of underlying thyroid disease and/or a clinical history rendering thyroid disease more likely. METHODS During a 1-month period, 156 consecutive inpatient TFTs on 143 adult patients (13 patients had multiple requests) were identi ed using the laboratory computer system and the patients clinical notes were recalled. In 12 cases the notes were not available, and two patients were excluded because the notes were incomplete. Consequently, 129 patient episodes (42 158

2 Thyroid function tests in acute 159 men, 87 women) were studied. The mean age was 69 years(median 74 years, range 23±95 years). This study was conducted in a 420-bed district general hospital serving a resident population of approximately The hospital has an accident and emergency unit and a medical admissions ward and, additionally, supports a subregional renal unit. TFTs were generally requested at, or soon after, the time of admission. The majority of patients were admitted to general medical wards; no patients were on the intensive treatment unit. Presenting symptoms are shown in Table 1. In all patients TSH was measured as a frontline test, with free thyroxine (ft 4 ) being reserved for cases in which the TSH concentration was abnormal. Hormones were measured using an Abbott AxSYM analytical system (Abbott Laboratories Ltd, Maidenhead, UK). The reference ranges for TSH and ft 4 were 0 49± 4 7 mu/l and 9 1±23 8 pmol/l, respectively. TABLE 1. Major presenting symptoms of 129 inpatients for whom thyroid function tests were requested Symptom Shortness of breath 43 Chest pain 35 Dizziness/unsteadiness/falls 31 Nausea/vomiting 18 Drowsiness/lethargy/tiredness/ weakness 17 Diarrhoea/constipation 17 Confusion/delirium 16 Palpitations 15 Fever/sweats/hot ushes 11 Collapse 8 Abdominal/epigastric pain 7 Loss of consciousness 7 Admitted for minor surgical procedure 15 Incidence (number of cases) A request for TFTs was considered to be justi ed if the clinical history increased the probability of thyroid disease contributing to the presenting complaint and/or if there were common symptoms or signs of thyroid disease (see Table 2). 4,6±8 A full drug history was recorded to establish whether drug interferences were a common problem in the interpretation of TFTs; in particular, drug treatment with amiodarone or lithium was considered to be a reasonable indication for requesting a TFT in an unwell patient. 8 For patients who were found to have abnormal TFTs, the clinical notes and laboratory computer system were reviewed again 1 year later to establish whether the thyroid abnormality had persisted following discharge. If no were available the patient s general practitioner was contacted directly with a request for further information and repeat investigations. For those cases in which either the patient died shortly after admission or direct contact with the general practitioner failed to elicit a further TFT request, the patient was classi ed as `lost to follow-up. RESULTS TFTs were felt to be clinically justi ed in 89 (69%) cases (see Fig. 1). Of these 89 cases, the TSH concentration was above the reference range in 17 (median 7 9 mu/l, range 4 8± 25 5 mu/l) and below the reference range in six (range 0 05±0 35 mu/l). There were no clear reasons for requesting TFTs in 40 (31%) cases; TSH concentrations were increased in ve of these patients (median 7 0 mu/l, range 5 0± 38 6 mu/l) but none had TSH concentrations below the reference range. The yield of abnormal results in the `clinically unjusti ed group did not differ signi cantly TABLE 2. Audit standard: justi able reasons for requesting a thyroid function test Clinical history Symptoms Signs thyroid disease Lethargy Goitre Neck or pituitary irradiation/surgery Constipation Myxoedema Turner s syndrome Weight loss/gain Ophthalmopathy Down s syndrome Cold/heat intolerance Tachycardia Addison s disease Refractory depression Atrial brillation Type 1 diabetes Anxiety state Bradycardia Drug treatment with amiodarone/lithium Dry skin Proximal muscle weakness Facial puf ness Carpal tunnel syndrome Hair loss Hyperemesis gravidarum Adapted from references 4, 6±8.

3 160 Lamb and Martin FIGURE 1. Justi able indications for requesting a thyroid function test in 89 inpatients. Note that some patients had multiple indications. from that in patients with clinical indications for thyroid investigation ( 2 with Yates correction 2 16, P=0 14). The clinical and biochemical features of patients with abnormal TFTs are summarized in Table 3. Of the 22 patients with increased TSH concentration, six had biochemical and/or clinical evidence of, none of whom were previously known to be hypothyroid. Five patients were known to have preexisting, and in three of these the dose of thyroxine was increased on the basis of the TFT results. Of the remaining 11 patients with raised TSH and no record of pre-existing thyroid disease, three had compensated with a

4 Thyroid function tests in acute 161 TABLE 3. Clinical features of patients in whom abnormal thyroid function tests were observed during their admission Patient Age/ sex Symptoms on admission TSH (mu/l) ft 4 (pmol/l) Diagnosis/ discharge summary Follow-up notes/ after discharge Explanation for initial abnormal TSH 1 32/M Chest pain < Thyrotoxicosis (re-started carbimazole) 2 34/F Constipation, nausea, vomiting 3 81/M Loss of consciousness, SOB 4 73/F Malaise, confusion Hyperemesis gravidarum (11 weeks pregnant), dehydrated Mild CRF, chest infection, poor diabetic control, AF Chest infection, (on T 4 ) Followed-up in OP clinic on carbimazole Normal delivery, patient well, no Died 5 days after admission, no 5 74/M SOB COPD, tachycardia, AF 2 months later: TSH /F SOB LVF, CVA Died 1 month later of bronchopneumonia, no 7 61/F SOB, chest pain, palpitations 8 89/F Off food, paranoia 9 66/M Cough with increasing SOB 10 80/F Chest pain, palpitations n/a AF, CCF 6 months later: TSH 3 2 (TSH before admission also normal) Dehydration, n/a chest infection, UTI AF, mild CRF, COPD, CCF, (started T 4 ) 6 months later: TSH SVT Repeated raised TSH/normal ft 4 on OP follow-up 11 84/F Chest pain, SOB Angina, (on T 4 ) Followed-up in OP clinic on T 4 thyrotoxicosis (?pregnancy) Compensated 12 67/F SOB, dizzy spells Oedema, CRF n/a 13 38/F Tired, appetite loss, weight loss Acute tubulointerstitial nephritis, essential hypertension 14 72/M SOB, chest pain, palpitations Ischaemic heart disease 15 66/M Palpitations Uncontrolled AF, recently started on amiodarone Repeated normal TSH on OP follow-up 4 months later: TSH year later stopped amiodarone: TSH /F Fall with cuts and bruising 17 80/F Chest pain, SOB Unstable angina GP follow-up 1 year later: TSH 9 2, ft Amiodarone effect CVA, AF n/a Compensated continued

5 162 Lamb and Martin TABLE 3. continued Patient Age/ sex Symptoms on admission TSH (mu/l) ft 4 (pmol/l) Diagnosis/ discharge summary Follow-up notes/ after discharge Explanation for initial abnormal TSH 18 78/F For surgery (hyperparathyroidism) 19 58/F Chest pain, palpitations, SOB, weight gain 20 73/F Wheezy, exercise intolerant 21 61/M Decreased consciousness, lethargy Parathyroid adenoma, Hashimoto s thyroiditis also seen on histology Hypothyroidism, TMA positive (titre 1:1600) (started T 4 ) AF, GP follow-up 1 year later: TSH 8 6, ft , TMA positive (titre 1:1600) Continues to be followed-up by GP on T 4 2 months later: TSH 10 7, ft (started T 4 ) CVA, CRF n/a, died 3 days after admission 22 87/F Falls, dizziness CRF, (on T 4, dose increased) 23 73/F Palpitations, dizzy, double vision 24 71/F Admitted for renal investigations 25 77/F Confusion, urinary incontinence 26 56/F Tiredness, lethargy, dry skin 27 91/F Increasing immobility, limb pain, SOB, sweats 28 65/F Nausea, sweats, confusion, collapse Continues to be followed-up as OP on T AF, CRF 1 year later: TSH 26 8, ft (started T 4 ) CRF, (on T 4 ) Hypothyroidism (on T 4, dose increased) Myelodysplastic syndrome, Hypothyroidism (on T 4, dose increased) Microcytic anaemia, OP on T 4 OP on T 4 OP (started T 4 ) OP on T 4 2 months later: TSH 26 8, ft (started T 4 ) Compensated An explanation for the initial abnormal thyroid function test (TFT) is based on information obtained from clinical notes and discharge summaries and from follow-up after discharge over a period of up to 1 year. All thyroid stimulating hormone (TSH) concentrations are expressed in mu/l and all ft 4 (free thyroxine) concentrations in pmol/l. For further details see text. AF=atrial brillation; CCF=congestive cardiac failure; COPD=chronic obstructive pulmonary disease; CRF=chronic renal failure; CVA=cerebrovascular accident; LVF=left ventricular failure; n/a=not available; OP=outpatient; SOB=shortness of breath; SVT=supraventricular tachycardia; TMA=thyroid microsomal antibody; T 4 =thyroxine; UTI=urinary tract infection. persistently raised serum TSH concentration after discharge. Three patients presumably had abnormal TFTs as a result of NTI since they had a normal TSH at follow-up. In one case the abnormal TFT result could be attributed to amiodarone, which the patient had recently started; TSH concentration measured several months after the patient had stopped taking amiodarone was normal. In four cases follow-up was not possible; one of these patients died within 3 days of admission. Of the six patients with reduced TSH concentrations, only one had a concentration below the detection limit of the assay and he was a known thyrotoxic patient being treated with carbimazole. Of the remaining ve, one was receiving replacement thyroxine therapy for and one patient s TSH may

6 Thyroid function tests in acute 163 have been low due to pregnancy, 5 but no postpartum follow-up results were available. One patient presumably had abnormal TFTs as a result of NTI since he had a normal serum TSH concentration after discharge. Two patients died during their admission before follow-up TFTs were taken. DISCUSSION Against a standard of the common clinical features of thyroid disease, the majority of inpatient TFT requests appeared justi able, although the abnormal results detected in this way rarely re ected the clinical indication (e.g. no patients with weight loss or atrial brillation were found to have thyrotoxicosis but a signi cant number of them did have increased TSH concentrations). Although NTI may have accounted for some of the abnormal TFTs observed in the seven patients who were lost to follow-up (three of whom died during their admission), it could only be con rmed as the explanation for one case of low serum TSH concentration and three cases of increased serum TSH concentration observed in this study. NTI did not appear to cause inappropriate management or interpretative dif culty in these cases, with TSH estimation being appropriately repeated when these patients were followed-up after discharge from hospital. In only one case did drug interference (amiodarone) appear to contribute to an abnormal TFT result. Our study supports the suggestion that TFTs may be abnormal in the hospital setting as a result of NTI, but the magnitude of this problem would appear to be lower than that observed in the earlier study of Spencer et al. 3 In contrast to that study, of the 12 patients with TSH concentrations in excess of 6 8 mu/l, 11 were found to have intrinsic thyroid disease. Although six patients had TSH concentrations below the reference range, in only one was it suppressed below the detection limit of the assay; this patient was a known thyrotoxic. The number of patients in this present study is considerably smaller than in the study of Spencer et al. 3 However, in their study, TFTs were not actually requested; instead, all samples passing through a hospital laboratory (and for which there was suf cient sample volume) were processed for TFTs. The population in that study included a high percentage of patients receiving glucocorticoids and/or with chronic renal failure. Additionally, the mean age of the patients was 44 years, considerably younger than in the present study and than the average age of acutely ill patients presenting to a general hospital. Consequently, their conclusions may not be transferable to routine clinical practice. Clearly, NTIs and drugs can affect TFT results. However, this needs to be balanced against the new cases of detected in this population and the cases of in which replacement treatment may have been inadequate (either as a result of non-compliance or too small a dose). In our experience, a signi cant number of TFT abnormalities observed in the setting of acutely ill hospitalized patients are consistent with underlying thyroid abnormality requiring investigation or treatment, particularly at higher concentrations of TSH ( 6 8 mu/l). We identi ed that abnormal thyroid function results were commonly not followed with appropriate further investigation and treatment, and this is a cause for concern. For example, evidence of inadequate replacement in patients with known was not always acted upon. In some cases, clearly abnormal TFTs were not documented in discharge summaries, leading to a signi cant delay in instituting appropriate further investigations and treatment. Patients with apparently compensated commonly did not have thyroid antibodies measured, 8 even though, in the hospital setting, the presence of thyroid antibodies appears to be a useful discriminator between intrinsic thyroid disease and NTI. 3 CONCLUSION Like all laboratory tests, TFTs should only be requested when clinically indicated. However, the manifestations of thyroid disease are so protean, and in many cases subtle, that requests may be justi ed in many unwell patients admitted to hospital. Indeed, in many of the patients in this study, it would have been remiss not to have requested a TFT. It is important that clinicians remain aware of the effects of NTI on TFTs and of the poor speci city of TSH in this situation. However, in clinical practice, thyroid dysfunction will often appear in the differential diagnosis of the acutely ill patient, particularly in the elderly. We would agree with Beckett and Wilkinson 1 that, faced with such potentially clinically important results, clinical biochemists

7 164 Lamb and Martin must assist clinicians in the interpretation and further investigation of these patients. Acknowledgements We are grateful to Mr E Kearney and the staff of the Clinical Biochemistry Department at Queen Elizabeth the Queen Mother Hospital, Margate, for measurement of the thyroid hormones. REFERENCES 1 Beckett G, Wilkinson E. Thyroid hormone metabolism and thyroid function tests in non-thyroidal. CPD Bull Clin Biochem 1998; 1: 9±14 2 Rae P, Farrar J, Beckett G, Toft A. Assessment of thyroid status in elderly people. BMJ 1993; 307: 177±80 3 Spencer C, Elgen A, Shen D, Duda M, Qualls S, Weiss S, et al. Speci city of sensitive assays of thyrotropin (TSH) used to screen for thyroid disease in hospitalized patients. Clin Chem 1987; 33: 1391±6 4 Anonymous. Sensitive thyrotropin measurements: utility and futility. Lancet 1989; i: Toft AD, Seth J. Sensitive thyrotrophin assays: excellent when properly used. BMJ 1987; 295: Weetman AP. Hypothyroidism: screening and subclinical disease. BMJ 1997; 314: 1175±8 7 Surks MI, Chopra IJ, Mariash CN, Nicoloff JT, Solomon DH. American Thyroid Association guidelines for use of laboratory tests in thyroid disorders. JAMA 1990; 263: 1529±32 8 Vanderpump MPJ, Ahlquist JAO, Franklyn JA, Clayton RN. Consensus statement for good practice and audit measures in the management of and hyperthyroidism. BMJ 1996; 313: 539±44 Accepted for publication 30 September 1999