Skip navigation

Signs and symptoms predict thyroid disease


On the face of it this headline has all the impact of "dog bites man". But one of the problems that clinical laboratories face, as they have for years, is the overwhelming tide of tests. When all they could offer was basic metabolic rate (and those old enough will remember how difficult they were to do), patients had full clinical workups before the test was done. A simple blood test is just that, simple, so anyone with any possibility of disease is investigated. All laboratories have a library of what they would regard as stupid reasons for requesting thyroid function tests - ingrowing toenail is one Bandolier remembers.

So some simple words from 20 years ago about the relationship between clinical signs and symptoms and the incidence of thyroid disease [1] are still relevant.

Study

Five-hundred consecutive inpatients and outpatients (those with known thyroid disease being excluded) to Flinders Medical Centre had case notes examined for:

  • thyroid function test (TFT) results.
  • clinical signs and symptoms noted by the clinician during the consultation from which the referral for TFT was made.
  • subsequent clinical diagnosis of thyroid status. The degree of clinical suspicion from signs and symptoms from a generally accepted (see box) list was correlated with final outcome.

Signs, symptoms and clinical suspicion of thyroid dysfunction

  1. Thyroid Goitre, thyroid bruit, fine tremor, weight loss, increased appetite, lid lag, sweating, heat intolerance, family history, lethargy, weight gain, hoarseness, dry skin, hair loss, cold intolerance, delayed reflex, constipation, short stature.
  2. Cardiovascular Recent myocardial infarction, chronic cardiac failure, coronary artery disease, arrhythmias, pulse >90/min, hypertension.
  3. Others Pneumonia, asthma, diabetes.

Degree of clinical suspicion

  • High Patient presenting with 5 or more signs/symptoms listed in groups 1 and 2.
  • Intermediate Patient presenting with 3 or 4 signs/symptoms listed in groups 1 and 2.
  • Low Patient presenting with 1 or 2 signs/symptoms listed in groups 1, 2 and 3.

Results

Of the 500 patients, 21 (4.2%) were found to have thyroid dysfunction needing treatment.

Degree of suspicion Number of patients Number with thyroid disease Percent with thyroid disease
High 23 18 78
Intermediate 35 1 2.9
Low 442 2 0.45
Total 500 21 4.2
In those patients with five or more clinical signs or symptoms in whom the degree of clinical suspicion was high, the majority (18, 78%) had thyroid disorder needing treatment. Using final diagnosis as the gold standard, the likelihood ratio for having a high degree of clinical suspicion was 82. This meant that from a pre-test probability of 4.2% the post-test probability was over 80%.
In those with high or intermediate degree of clinical suspicion, 19 (33%) had thyroid disorder needing treatment. Using final diagnosis as the gold standard, the likelihood ratio for having a high or intermediate degree of clinical suspicion was 11. This meant that from a pre-test probability of 4.2% the post-test probability approached 40%.



In those with low degree of clinical suspicion, 2 of 442 patients (0.45%) had thyroid disorder needing treatment. Using final diagnosis as the gold standard, the likelihood ratio for having a low degree of clinical suspicion was 0.1 (calculated as the likelihood ratio of a negative test). This meant that from a pre-test probability of 4.2% the post-test probability was less than 1%.

Comment

"Look at the relative size of the likelihood ratios for a brief, immediate, relatively cheap history and a much longer, delayed, and relatively expensive exercise electrocardiogram. There is no contest. Likelihood ratios for key points in the history and physical examination, both for this and for most other target disorders, are mammoth and dwarf those derived from most excursions through high technology."[2]


Bandolier has always found these words comparing clinical examination and exercise testing compelling, but has struggled to find examples. Thyroid function testing appears to be one. Of course, these are "old" data, but Australian patients being examined in 1977 won't be that different from people in the UK in 1997.

Since the incidence of thyroid disease in the general population is the order of 1% or below, some selection has already gone on to create a 4.2% incidence in the population in the paper. But that serves only to emphasise the very low yield of thyroid disease (0.45%) in those patients in whom there was a low clinical suspicion. It prompts the question, why do a TFT? But post-test probabilities of more than 30% for at least 3 signs and symptoms, and of more than 80% for at least 5 signs and symptoms would seem to make TFTs worthwhile in confirming the diagnosis.

The authors discuss the "heavy economic and logistic burden of biochemical screening" that biochemical screening was causing. Plus ça change.

References:

  1. GH White, RN Walmsley. can the initial clinical assessment of thyroid function be improved? Lancet 1978 ii: 933-5.
  2. Clinical Epidemiology (Eds DL Sackett, RB Haynes, GH Guyatt, P Tugwell), 2nd Edition, 1991, p132.



previous or next story in this issue