Goiter

Prior Probability

The prior probability of a goiter is affected by many variables, including the patient's body surface area, sex, and regional variations associated with the endemic iodine deficiency. Two recent European studies of thyroid volume among community samples of healthy adults give us insight into the prevalence of goiter in the non–iodine-deficient area: 4% of patients in Spain (95% confidence interval [CI], 3%-6%)⁸ and 10% of patients in France (95% CI, 9%-11%)⁹ had palpable goiters. Unfortunately, the thyroid volume was not confirmed for patients with palpable goiters. Nonetheless, we can make some inferences that give us good starting points. The WHO defines an iodine-deficient area by the prevalence of goiter in school-aged children. According to normative population values, children who live in a non–iodine-deficient area should have a goiter prevalence of less than 5%.¹ Adults might have palpable thyroid glands for reasons other than iodine deficiency, so prevalence values slightly higher make sense. A starting point of 5% to 10% for healthy adults makes sense for the prior probability of a palpable thyroid.

Population for Whom a Goiter Disease Should Be Considered

• Symptoms of hyperthyroidism or hypothyroidism

• Children, especially those in endemic iodine deficiency locales

• Pregnant and lactating women

• Elderly patients

• Patients with excessive radiation exposure

• Patients with Down syndrome

Detecting the Likelihood of a Goiter

Because examining children is different from examining pregnant women for thyroid disease, we cannot combine the data (see Table 21-10). The techniques for examination, however, are similar. We have no data for the results of thyroid palpation in non-pregnant adults because epidemiologic studies of normal adults’ thyroid volume exclude those with palpable enlargement.

Palpating thyroid tissue in both lobes of a volume greater than the volume of the patient's distal thumb phalanx increases the likelihood of a goiter, but there will be false-positive results.

Reference Standard Tests

Ultrasonography.

In epidemiologic research, urinary iodine studies are evaluated along with thyroid palpation.

Clinical Scenarios

How Large Are These Thyroid Glands?

For each of the following patients, assessment of thyroid size is an important part of the clinical examination. In case 1, a 32-year-old woman presents with symptoms and findings consistent with hyperthyroidism, but she has no exophthalmos and has always been anxious. In case 2, a 55-year-old man has a diagnosis of Graves disease, and the choice is made for radioactive iodine ablation therapy. In case 3, a 64-year-old man has a goiter that causes discomfort on swallowing, and thyroxine is administered in an attempt to shrink the thyroid gland.

Why Assess the Thyroid Gland for Size?

A goiter is simply an enlargement of the thyroid gland and may result from hormonal or immunologic stimulation of gland growth or the presence of inflammatory, proliferative, infiltrative, or metabolic disorders (Table 21-1). A common error among clinicians first learning about the thyroid is to associate thyroid size with function; a goiter, however, can be present in hyperthyroidism, hypothyroidism, or a euthyroid state. Determining whether a thyroid is enlarged can aid in diagnosis, differential diagnosis, and decisions about laboratory testing; in determining specific therapy and therapeutic dosing; and subsequently in monitoring of the clinical course. For example, when a patient presents with symptoms that could be caused by hyperthyroidism, the detection of a goiter increases the likelihood that thyrotoxicosis is present.² If the patient described in the first case had an enlarged thyroid, hyperthyroidism would be a likely diagnosis.² On the other hand, if her gland were of normal size, anxiety might be the explanation for her symptoms. Determination of thyroid size also is useful once a specific disease is diagnosed. In patients with Graves disease, for example, thyroid size may be a factor in determining choice of treatment because patients with smaller glands are more likely to go into immunologic remission during antithyroid drug therapy.³ If radioiodine is the chosen treatment, as in the second case, the size of the gland is often used in calculating the dose to be administered.⁴ Finally, responses to various therapies can be monitored clinically by assessing thyroid size, such as the attempt to shrink a large goiter with thyroid hormone administration in the third case.⁵

The Anatomic Basis of Thyroid Examination

Landmarks and Relation to Other Structures

The thyroid gland is located in the anterior neck and usually consists of 2 lobes connected at their lower midregions by a transverse isthmus (Figure 21-1). The most prominent structure in the anterior neck is the thyroid cartilage. Inferior to the thyroid cartilage lies the cricoid cartilage, and inferior to this lies the isthmus of the thyroid gland, which can be as low as the level of the fourth tracheal ring. Each thyroid lobe lies against the sides of the trachea, extending up from the isthmus to the region of the cricoid and thyroid cartilages and downward toward the clavicles. The posterior portion of each lobe lies beneath the belly of the ipsilateral sternocleidomastoid muscle. Because the fascial envelope of the thyroid gland is continuous with the pretracheal fascia of the cricoid cartilage and hyoid bone, the thyroid ascends and descends with the laryngeal structures during swallowing.

How Large Is the Normal Thyroid?

The normal thyroid size for a population is largely determined by the supply of iodine in the diet, with a tendency to larger glands in iodine-deficient areas.^{6, 7, and 8} Consequently, studies of clinically normal thyroid glands have demonstrated sizes that span an extreme range in euthyroid individuals, differing by geographic location and varying through time within a given region as iodine supplementation has been instituted. Until the middle of this century, most authors considered a typical thyroid gland to be about 20-25 g, and a commonly accepted upper normal size was 35 g.^{8, 9, 10, and 11} More recent studies in iodine-supplemented populations have reported mean weights of 10 g or less and an upper normal size of 20 g.^{12, 13} Although a value of 35 g may still apply in iodine-deficient areas, an upper normal weight of 20 g is probably appropriate for most parts of the western world and will be used for this analysis. With this definition, the prevalence of goiter is typically 2% to 5% in iodine-replete regions.^{13, 14}

How to Examine the Thyroid Gland to Determine Size

The normal thyroid is rarely visible because of its relatively small size, partial concealment by the sternocleidomastoids, and soft texture, and it may be marginally palpable.^{5, 9, 15} Enlargement is initially observed as an increase in the size of the lateral lobes to palpation.^{5, 8} Further growth results in a gland visible in the anterior side of the neck that can be seen when inspecting from the side^{16, 17} and from the front with the patient's neck extended.^{5, 7, 15, 18} With increasing size, the gland becomes even more prominent on inspection from the side and it becomes visible from the front with the patient's head in a normal position. Ultimately, a large goiter is easily palpable, has prominence from the side of greater than 1 cm, and is visible from the front at a distance.^{5, 17, 18}

As a result of observations on these patterns of enlargement, various systems have been described to size a thyroid gland according to (1) the estimated weight^{19, 20, and 21}; (2) the volume relative to the size of normal glands^{5, 8}; (3) the presence or absence of palpable or visible enlargement^{8, 18, 22}; (4) the degree of visible prominence when the neck is viewed laterally¹⁷; (5) neck circumference determined by tape measure^{23, 24}; (6) the surface area of the gland projected onto the skin^{22, 25}; and (7) the maximum width of the lower poles, measured with a ruler or calipers.²⁶ Many of these rating scales were developed for epidemiologic studies of goiter in endemic areas and were intended to classify significant goiters rapidly (with examination time in some studies averaging only 18 seconds per subject).⁵ As a result, many are of little use for the smaller thyroid glands observed in regions without significant levels of endemic goiter. Most studies from which data for accuracy and precision of goiter determination can be derived do not report specifics of thyroid examination technique. Consequently, there is no objective evidence to support the use of one examination method over another.^{23, 24, and 25, 27, 28} Many of the variations are minor, so shared features will be described.

The patient should be comfortably positioned, either standing or seated, with the neck in a neutral position or slightly extended. The region of the neck below the thyroid or cricoid cartilage should be observed from the front, with good cross-lighting to accentuate shadows and highlight masses. If an abnormality is suspected, the neck should be moved as appropriate to alter the prominence of the area under suspicion. A particularly useful maneuver is inspection during full extension of the patient's neck. This position stretches superficial tissues over the thyroid gland, which is pressed against the relatively unyielding trachea, and enhances visibility of the gland. Inspection of the neck from the side, looking for a prominence protruding from the normally smooth and straight contour between the cricoid cartilage and the suprasternal notch, can reveal enlargement.¹⁷ The amount of prominence should be measured with a ruler (Figure 21-2). This method requires a certain degree of guesswork in deducing where the normal neck contour would lie, but the measurement can provide information useful for ruling in the presence of a goiter. There is no particular spot to place the ruler; it merely serves as a visual guide in estimating the degree of protrusion.

After inspection, the gland is palpated, and this is where the greatest differences in methods arise. Clinician preference varies about palpation with fingers or thumbs, an approach from the front or from behind the patient, and whether each lobe is palpated by the ipsilateral hand or the opposite hand. In the absence of data to support a specific method, though, examiners should use the approach with which they are most comfortable. Regardless of the technique used, it is often useful to first attempt to locate the thyroid isthmus by palpating between the cricoid cartilage and suprasternal notch. An isthmus may not be felt, but if it is, this can help locate the gland. When palpating the lobes, it is beneficial to relax the sternocleidomastoids. To better feel the left lobe, for example, the neck can be slightly flexed and rotated to the left to relax the left sternocleidomastoid and to make space for the palpating fingers or thumb between the sternocleidomastoid and trachea. There are certain additional maneuvers that may be useful, such as measuring neck circumference or the dimensions of a lobe with calipers, but no information is available to assess accuracy or precision of these techniques. Other elements of the thyroid examination that are carried out concomitantly with size assessment include determining gland texture, gland mobility, tenderness, and the presence of nodularity. Auscultation also may be performed for the presence of bruits. These features have their own implications but are not central to determining the presence of a goiter and so are beyond the scope of this discussion. If no thyroid is detected in the neck, it may be maldescended or intrathoracic. Methods of examining for these variants will not be discussed here, because, again, no information is available to analyze the reported techniques.

Dogma holds that the thyroid examination is improved by having the patient swallow during both inspection and palpation. Indeed, it has been stated that swallowing increases sensitivity of inspection alone to that of inspection combined with palpation.²⁸ No study, however, has actually analyzed whether a swallowing maneuver is of benefit, although most examiners believe it is. The movement resulting from swallowing accomplishes several things. First, it changes the shadowing of any mass, enhancing visual detection of a bulge in the neck contour that may be too subtle to be detected otherwise. Second, movement of the thyroid raises a low-placed gland up from below the sternal notch or lower sternocleidomastoid, making it accessible when it may not have been so previously. Third, as in any palpation technique, movement of the object against the palpating hand increases definition. Finally, because only the larynx, upper trachea, and thyroid gland move with swallowing, this maneuver can aid in anatomic localization.²⁹ The degree of excursion of the thyroid on swallowing is proportional to the size of the bolus swallowed, so the patient should be given a sip of water.³⁰

When the thyroid is examined to determine the presence of a goiter, the goal is to estimate gland size. Most endocrinologists express findings in absolute mass or as relative to an upper limit of a normal-sized gland, such as “normal” or “2 to 3 times normal size.” Many nonendocrinologists have some difficulty quantifying thyroid mass, but this ability is crucial in accurately classifying a gland, as will be discussed in the analysis of accuracy.

False-Positive and False-Negative Goiter Results

Finding a goiter when one is not present may simply be an error in detection. There are, however, several common causes of a false-positive goiter or pseudogoiter finding. One is simply an easily palpable gland in a thin individual.⁵ Because the entire thyroid is so accessible, the tendency is to interpret this accessibility as being due to an enlarged gland rather than the true reason, a decrease in interfering tissues that normally block access to the gland. A second cause is a variant of the normal placement of the thyroid gland in the neck. In some individuals, the gland is higher than usual, and this prominence is again attributed to enlargement.³¹ A third anatomic variant has been termed Modigliani syndrome.³² In Modigliani syndrome, the thyroid actually lies in a normal position below the cricoid cartilage, but such individuals possess long, curving necks that enhance the prominence and palpability of the gland. A fourth condition producing pseudogoiter is a fat pad in the anterior and lateral portion of the neck.²⁴ Although this condition may be more common in obese individuals, it can also be found in those of normal weight, particularly young women. With experience, examiners can learn to differentiate this from true thyroid tissue by the differing textures and shapes and the lack of movement of a fat pad with swallowing. Another cause involves the thyroid being pushed forward by lesions behind it, making it more easily palpable.^{5, 33} Finally, any enlargement in the vicinity of the thyroid gland may be mistaken for an enlarged thyroid gland, particularly if it is adherent to the thyroid or larynx and so moves with swallowing.²⁹

There are 3 principal causes of false-negative goiter detection in addition to true misclassification. The first and probably most common cause, of course, is an inadequate physical examination. In some circumstances, an imperfect examination is unavoidable, as when a patient is intubated. In most cases, however, with a little effort, a good examination can be performed on virtually all patients. Second, some individuals, particularly the obese, the elderly, or those with chronic pulmonary disease, have short and thick necks, obscuring the thyroid.^{5, 24, 34} Some patients also have an atypical thyroid placement, such as a retrosternal location, or lobes that are lateral and obscured by the sternocleidomastoids, making palpation difficult.³⁵

Precision of Estimating Thyroid Size

Interobserver Variability

Data on interobserver precision in estimating thyroid size are available both for rating scales that attempted to place glands in one of 3 or 4 categories according to palpability and visibility and for simple estimation of the presence or absence of a goiter (Table 21-2). Agreements were good to very good in both cases. When glands were placed in categories, κ ranged from 0.47 to 0.74, with a value from combined data of 0.70 (95% confidence interval [CI], 0.68-0.72).^{7, 8, 36, 37} (The κ statistic and other statistical measures are defined in the introductory article to this series.³⁸) For determination of goiter, κ ranged in these 4 studies from 0.38 to 0.77, with a value for combined data of 0.77 (95% CI, 0.76-0.79). Similar results were reported in another study,³⁹ in which observers determined whether individual lobes were enlarged, with κ from 0.32 to 0.62, and in yet another report⁴⁰ that determined the presence of a goiter, with κ from 0.10 to 0.54. Because of the nature of the rating scales used in 2 of these studies,^{8, 37} we can specifically compare interobserver variability for the techniques of inspection (κ = 0.65; 95% CI, 0.62-0.69) and palpation (κ = 0.74; 95% CI, 0.67-0.82). These techniques did not differ significantly in the level of agreement, and both were very good (Table 21-3).

As might be expected, most disagreements between observers involved smaller glands and those near the cutoff for goiter determination, and most disagreed by only 1 stage in classifications.^{7, 8, 36, 37} Agreement may be better between examiners with greater experience than between those with differing levels of training.⁴⁰

Intraobserver Variability

In 2 studies,^{6, 22} examiners placed thyroid size in categories of enlargement and repeated the examination on a separate occasion (Table 21-4). These data produced a κ from combined numbers of 0.59 (95% CI, 0.52-0.65) for placement in all categories of the rating scales used by the examiners. For simply determining the presence or absence of goiter, κ ranged from 0.47 to 0.79, with a κ from combined data of 0.65 (95% CI, 0.63-0.67), which is very good. Similar results were reported in a study of patients with various thyroid diseases, in which κ ranged from 0.54 to 0.74.³⁹ Intraobserver agreement was slightly better for the inspection component of the examination (κ = 0.73; 95% CI, 0.71-0.76) than for palpation (κ = 0.65; 95% CI, 0.63-0.67) (Table 21-5).

Accuracy of Estimating Thyroid Size

Three criterion standards have been used in assessing the accuracy of thyroid size determination: weight measured after surgical or postmortem removal, ultrasonographic assessment, and nuclear scintigraphy. Ultrasonographic assessments of thyroid weight correlate well with true gland weight as determined after excision (r = 0.88-1.0), although there is lack of agreement as to the best formula to use for estimating size.^{18, 21, 41} Nuclear scan determination is a little less reliable but acceptable (r = 0.77-0.98).^{9, 42, 43} Again, different formulas have been used to translate the scintigraphic profile to thyroid volume.^{9, 42, 43}

Combining data from 9 studies of detection of goiter by physical examination,^{12, 17, 18, 21, 44, 45, 46, 47, and 48} the sensitivity from combined data was 0.70 (95% CI, 0.68-0.73) with a specificity of 0.82 (95% CI, 0.79-0.85) (Table 21-6). If a goiter was clinically detected, the positive likelihood ratio (LR+) of one being present was 3.8 (95% CI, 3.3-4.5). Conversely, if a goiter was not thought to be clinically present, the negative likelihood ratio was 0.37 (95% CI, 0.33-0.40). These likelihoods are comparable with or better than those for many other physical signs^{49, 50} and were not affected by the presence of single or multiple nodules.⁴⁸ Experienced examiners were somewhat more accurate in their assessments than more junior colleagues.⁴⁸

Some authors have defined specific stages of thyroid enlargement according to the usual sequence of changes that occur as the thyroid gland increases in size. Because some of these staging classifications incorporate observations not normally used in simply estimating thyroid mass, they can significantly enhance the predictive abilities of the clinician (Table 21-7). In the combined data from 4 studies,^{19, 20, and 21, 48} when a clinician thought that a thyroid gland was of normal size, the LR+ of goiter being present was 0.15 (95% CI, 0.10-0.21). If classified as 1 to 2 times normal size, the LR+ was 1.9 (95% CI, 1.1-3.0), and for greater than 2 times normal, the LR+ was 25 (95% CI, 3.6-175).

Certain staging methods for thyroid enlargement can help clarify the true status of some of the patients with glands thought to be 1 to 2 times normal size after routine inspection and palpation.^{14, 17} The amount of prominence of the thyroid on lateral inspection, for example, resulted in a high likelihood of goiter if it was greater than 2 mm (Table 21-8). Of further utility was finding that a gland was not visible with the neck extended, a result that effectively ruled out a goiter.

Bias in Estimating Thyroid Size

When the results from 4 studies^{19, 20, and 21, 48} estimating thyroid gland weights were combined, a regression line was produced describing the bias in gland size determination (Figure 21-3). This clearly shows that sizes of smaller glands are routinely overestimated, whereas those of larger glands are underestimated. The size at which this crossover occurs corresponds to about 2 times normal size. The practical application of this finding is that glands in the 1- to 2-times-normal-size category fall in the range in which size is typically overestimated.

The Bottom Line

To determine whether a goiter is present, follow these steps:

1. Examine the thyroid gland by inspection and palpation.

2. Categorize thyroid size as normal or goiter. Subcategorize goiter as small goiter (1-2 times normal) or large goiter (greater than 2 times normal).

3. If you placed the thyroid in the small-goiter category, consider whether you overestimated the size; determine whether there is any prominence in the profile of the neck in the region of the thyroid when viewed laterally (classify the prominence as ≥2 or >2 mm), and determine whether the gland is not visible from the front with the neck extended.

4. Place your patient in one of the following categories: “goiter ruled out,” normal thyroid size or thyroid considered to be not visible with neck extended; “goiter ruled in,” large goiter present or lateral prominence greater than 2 mm; or “inconclusive,” all other findings.

Author Affiliations at the Time of the Original Publication

Division of Endocrinology and Metabolism, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada

Acknowledgment

This work was supported by the Alberta Heritage Foundation for Medical Research.

References

Clinical Scenario

A 34-year-old woman had a child about 14 months ago. She had been breast-feeding her newborn but stopped about 2 months before her routine visit with you. She complains that her weight has not gone back to baseline and that her skirts are tight at the waist and her blouses are tight at the neck. Does she simply need to lose weight, or could she have a goiter?

Updated Summary on Goiters

Original Review

Updated Literature Search

Our updated literature search used the parent search strategy for The Rational Clinical Examination series, combined with the subject headings “exp Goiter,” “limited to diagnosis,” “radionuclide imaging,” “epidemiology,” and “ultrasound studies,” published in English from 1994 to 2004. We also crossed the clinical subject headings with “meta-analysis,” “ROC curve,” and the textword “systematic review” in both MEDLINE and the Cochrane databases. The results yielded 135 titles, for which we reviewed the titles and abstracts; 10 were selected for additional review. Two additional articles were selected for review from the references. We included articles that allowed us to calculate the sensitivity and specificity, or the observer variability, of the clinical examination for a goiter.

Improvements in the Data Presented in the Original Publication

The actual techniques for palpating the thyroid are described well in the original publication. However, 3 methods for assessing thyroid size from palpation were presented: estimates of thyroid volume (in grams), lateral prominence of the thyroid (in millimeters), and a 5-level ordinal assessment based on palpability and visibility. The World Health Organization (WHO) proposed a simplified classification for the presence of a goiter: an individual has a goiter when each lateral lobe has a volume greater than the individual's terminal phalanx of the thumb.¹ A grade 1 goiter will be palpably enlarged but not visible when the neck is in a normal position. A grade 2 goiter will be palpably enlarged and visible with the neck in the normal position. Most of the work on establishing these criteria comes from epidemiologic studies of endemic iodine deficiency that used children as the study subjects. The epidemiologic studies use examiners with considerable thyroid examination experience.

The effect of changing the threshold for the clinical screening test (palpation) changes the performance of the test. The interobserver variability when performed by experienced examiners is acceptable with both the 1960 and 1994 criteria. It is important to compare the size of the thyroid to the thumb because the case definition is not “any” palpable thyroid but one that is larger than the distal thumb. One study in a high-prevalence area found that defining a palpable thyroid as enlarged, without comparing the size to the thumb, increased the clinical goiter rate by 20%.²

Changes in the Reference Standard

The reference standard for thyroid enlargement remains ultrasonography. A goiter is defined as a thyroid gland of increased volume. However, the appropriate threshold for identifying the patient as having enlargement vs not having enlargement is evolving. WHO recognizes endemic iodine deficiency as a global health problem. The prevalence rate of goiters in school-age children defines regions as having endemic iodine deficiency vs normal iodine status. The definitions of normality for children may be different from those of adults because thyroid volume depends on body surface area (it also varies by sex). Areas of endemic iodine deficiency may be severely affected by malnourishment, and this in turn affects the size of thyroid glands.

The older 1960 WHO standard for thyromegaly required that a lobe of the thyroid have greater volume than the terminal phalanx of the child's thumb. These criteria, established before ultrasonography, defined endemic iodine deficiency as a population with a greater than 10% prevalence of goiter. Palpation was the only method for assessing thyroid volume, and the use of the child's thumb as a comparative standard would seemingly account for both the child's sex and body surface area. In 1994, a newer threshold was proposed for epidemiologic research that used ultrasonography and normative thyroid volume adjusted for body surface area. The newer threshold decreased the prevalence level to more than 5% to define iodine deficiency areas but also simplified the clinical criteria for a goiter. A key question is whether a universal normative standard should be used for thyroid volume (eg, a universal threshold volume above which defines a goiter, or above a percentile for the universe of patients) or whether local reference standards should be established (eg, thresholds developed within a defined geographic region).³

Results of Literature Review

When palpating the thyroid, compare the results of each lobe to the subject's distal thumb. A thyroid with both lobes larger than the patient's distal thumb is considered palpably enlarged (see Table 21-9).

Evidence From Guidelines

The US Preventive Health Services Task Force evaluated the role of thyroid-stimulating hormone (TSH) screening in healthy adults and observed that men have a lower prevalence of unrecognized and unsuspected thyroid disease compared with women.⁶ High risk patients for thyroid disorder include the elderly, postpartum women, those with high levels of radiation exposure (>20 mGy), and patients with Down syndrome. However, the task force concluded that the data are inconclusive for recommending TSH screening. The task force does not address clinical screening with palpation. Health Canada guidelines came to similar conclusions as the US Health Services Task Force.⁷

Clinical Scenario—Resolution

Returning to prenatal weight is a postpartum problem for many women. This patient has an unusual complaint of clothing feeling tight around the neck, so you feel obligated to palpate for a thyroid. Despite the enlargement, many patients do not recognize that they have a goiter. Goiters are more common in women, especially during pregnancy and in lactating mothers. When you palpate her thyroid, you need to use proper technique and make sure that any palpable thyroid tissue moves upward when she swallows. If you feel thyroid tissue, decide whether the volume of the palpable tissue in both lobes is greater than the volume of her distal thumb. Although this approach of assessing volume has been validated in children and not in adults, inexperienced examiners may have difficulty deciding whether the thyroid volume is normal compared to endocrinologists who assess the size compared with a normal gland (eg, 1.5 times normal or 2 times normal). If you are uncertain whether the gland is normal, ultrasonography would confirm the presence or absence of a goiter. You should also assess more fully for signs and symptoms of thyroid dysfunction. Although the most common cause for inability to lose weight postpartum may be lack of exercise, a sensitive TSH assay would be required to make sure she does not have hypothyroidism.

References for the Update

Evidence Summary and Review 1

Description of Tests and Diagnostic Standard

Thyroid size by the WHO 2001 criteria: grade 0, normal; grade 1, both lobes larger than the distal phalanx of the thumb, but the gland is not visible; grade 2, both lobes palpably enlarged but also visible. The examination techniques are well described and the examiners had their reliability confirmed.

Thyroid size was confirmed by ultrasonography. Patients also provided urine samples for urinary iodine.

Main Outcome Measure

Thyroid size. Values below the 90th percentile for the regions were considered not enlarged.

Main Results

The 2 endocrinologist examiners had a κ of 0.70 for their agreement on the scoring scheme. The criteria have good discriminative properties for identifying patients with goiter (see Table 21-11).

Conclusions

Level of Evidence

Level 1.

Strengths

Study population included sample of low-risk and higher-risk patients. The examiners had their reliability assessed. The examination techniques are well described.

Limitations

Examination done among pregnant patients only. The reference standard was a threshold that some might consider too low; a value greater than the 90th percentile was considered as goitrogenous.

Palpating pregnant women's thyroid glands may be easier than palpating the thyroid gland of nongravid subjects. However, it is possible that the simpler grading scheme and training of the examiners led to excellent reliability. The low negative likelihood ratio is impressive, suggesting that the finding of a nonpalpable gland during pregnancy rules out thyromegaly. Of course, patients can have substernal goiters, so we know that there will be some false-negative results (but not many).

Because the lower threshold for defining a goiter was used (the 1960 WHO criteria of 10th percentile rather than the currently recommended 5th percentile), we would expect the specificity and the positive likelihood ratio to be worse. However, compared with the results for using a 5-level scheme, the results are promising.

Evidence Summary and Review 2

Description of Tests and Diagnostic Standard

Three examiners examined each child. One of the examiners was described as “experienced,” 1 was an experienced paramedic, and 1 was an inexperienced “expatriate physician.” Each examiner evaluated each child in the morning and then again in the afternoon. Although they were not given their morning results, the examiners were not blinded to the child. The ultrasonographic testing for each child was also repeated.

Main Outcome Measures

Interobserver and intraobserver variability.

Main Results

Seventy-five percent of the children had goiter by the reference standard ultrasonogram.

The inexperienced physician had a low intraobserver variability with both the 1960 and 1994 criteria (κ of 0.36 and 0.44, respectively). The intraobserver variability for the experienced examiners was similar for both criteria (κ of 0.57-0.58 for the 1960 criteria and 0.53-0.60 for the 1994 criteria). The performance of the inexperienced observer improved over time (κ of 0.26 during the first 3 days of the study compared with 0.56 for the last 3 days, using the 1960 criteria).

The ultrasonographer had an intraobserver variability that resulted in a reclassification of 14% of patients from morning to afternoon examinations (κ of 0.63).

Conclusions

Level of Evidence

Level 3.

Strengths

Three examiners of various levels of experience. Intraobserver variability was assessed.

Limitations

The ultrasonographer had poor precision, making the quality of the reference standard doubtful.¹ There was a lack of independence in the examination by each clinician. The study was done in an exceedingly-high-prevalence area.

We include a review of this article for several reasons. First, it seems clear that the intraobserver variability is better for experienced examiners. Second, this study demonstrated that the inexperienced observer's precision improved during the course of the study, which allows us to infer that practice is helpful. Third, it is important to apply the clinical criteria as they are currently specified. “Any” palpable enlargement does not qualify the patient has having a goiter, because each lobe must be of greater volume than the distal phalanx of the thumb. Finally, the reliability of the ultrasonographic reference standard was probably too low. For this reason, we do not show the sensitivity and specificity of the individual examiners.

Reference for the Evidence

Evidence Summary and Review 3

Description of Tests and Diagnostic Standard

Two examiners evaluated each child, blinded to each other's result. An ultrasonographer evaluated each subject, blinded to the clinicians’ examinations. The examiners and ultrasonographer were all experienced in goiter epidemiologic studies. The clinicians recorded their findings according to the WHO 1960 criteria for goiter and the 1994 criteria (Table 21-12). The WHO upper limit of the thyroid volume, adjusted by the subject's sex and body surface, was used as the reference standard.

Main Outcome Measures

Main Results

In the community with mild iodine deficiency, κ was 0.47 between examiners for the 1960 vs 1994 criteria and 0.53 for the 1994 criteria. In the severe iodine deficiency site, κ was 0.67 between examiners for both the 1960 and 1994 criteria.

In the high-prevalence village, the accuracy of the clinical examination was similar for the 1960 and 1994 criteria (see Table 21-13). In the low-prevalence village, the clinicians estimated a prevalence of 20% to 21% with the 1960 criteria and 25% to 26% with the 1994 criteria; however, the actual prevalence was 12%.

Conclusions

Level of Evidence

Level 1.

Strengths

Large sample size of a community-based population for whom it was reasonable to screen for thyroid disease. The study subjects were not enrolled because of a suspicion for disease.

Limitations

The sampling frame is not specified and the study enrolled no adults. All the clinicians were experienced examiners, which limits generalizability.

The 1994 revised criteria for goiter simplified the scale from 5 to 3 levels. Overall, the performance between the 2 criteria appears similar. With fewer choices, the 1994 criteria ought to be more reliable, especially when used by less experienced examiners. The newer criteria required only that the thyroid be palpable to be considered clinically enlarged. In areas of low prevalence, this would lead to overestimates of ultrasonographically proven thyromegaly. A 2001 revision of the criteria clarified that “palpable” meant an enlargement of both lobes to a volume greater than the distal phalanx of the subject's thumb.

It is disappointing that the experienced thyroid examiners did not have higher diagnostic accuracy. Other studies have found that the clinical examination overestimates the volume of the thyroid in schoolchildren.² A partial explanation may relate to problems with using a worldwide WHO fixed cut point for ultrasonographic size rather than with using local references values adjusted for sex and age.³

References for the Evidence

Clinical Scenario

A 34-year-old woman had a child about 14 months ago. She had been breast-feeding her newborn but stopped about 2 months before her routine visit with you. She complains that her weight has not gone back to baseline and that her skirts are tight at the waist and her blouses are tight at the neck. Does she simply need to lose weight, or could she have a goiter?

Clinical Scenario—Resolution

Returning to prenatal weight is a postpartum problem for many women. This patient has an unusual complaint of clothing feeling tight around the neck, so you feel obligated to palpate for a thyroid. Despite the enlargement, many patients do not recognize that they have a goiter. Goiters are more common in women, especially during pregnancy and in lactating mothers. When you palpate her thyroid, you need to use proper technique and make sure that any palpable thyroid tissue moves upward when she swallows. If you feel thyroid tissue, decide whether the volume of the palpable tissue in both lobes is greater than the volume of her distal thumb. Although this approach of assessing volume has been validated in children and not in adults, inexperienced examiners may have difficulty deciding whether the thyroid volume is normal compared to endocrinologists who assess the size compared with a normal gland (eg, 1.5 times normal or 2 times normal). If you are uncertain whether the gland is normal, ultrasonography would confirm the presence or absence of a goiter. You should also assess more fully for signs and symptoms of thyroid dysfunction. Although the most common cause for inability to lose weight postpartum may be lack of exercise, a sensitive TSH assay would be required to make sure she does not have hypothyroidism.