Hemithyroidectomy for non-toxic nodular goitre

Background: Auto-immune thyroiditis, associated with detectable thyroid auto-antibodies (TAAs) and lymphocytic infiltration into the gland, is known to be associated with progressive development of hypothyroidism. This study examines those patients who required surgical treatment of non-toxic nodular goitre by hemithyroidectomy to determine whether the presence of TAAs in the circulation and/or lymphocytic infiltration of the gland resulted in a significant degree of post-operative hypothyroidism. Method: This was a prospective study, with data collected retrospectively. All patients operated on for thyroid disease in the unit over a 5-year period were documented, and those patients treated by hemithyroidectomy for non-toxic nodular goitre formed the study group. TAAs were measured, histology re-examined and patients followed up for at least 2 years to detect clinical or sub-clinical hypothyroidism. Results: Of the 158 patients who comprised the study group, 38 (24.1%) developed hypothyroidism. Forty-one (25.9%) had circulating levels of TAAs, 31 of these (75.6%) having high levels. Of these 41 patients, 22 (53.7%) became hypothyroid. Twenty (64.5%) of the 31 patients with high levels of TAAs required post-operative thyroxine. Of the remaining 117 patients with no detectable TAAs, only 16 (13.7%) became hypothyroid. Conclusion: All patients requiring treatment by hemithyroidectomy should have circulating TAA measurements carried out pre-operatively. If positive, they should be followed up indefinitely because of the strong possibility (P<0.001) of the development of hypothyroidism.

Key words: Hemithyroidectomy, hypothyroidism, lymphocytic infiltration, nodular goitre, thyroid auto-antibodies

Auto-immune thyroiditis, Hashimoto’s ‘struma lymphomatosa’, has long been recognised to result in the progressive development of hypothyroidism.^1-5 The diagnosis is usually made by the detection of circulating thyroid anti-bodies (TAAs) in the serum of affected individuals, and is associated with a diffuse lymphocytic infiltration into the gland.^3,5 Anti-thyroid microsomal antibodies (TMAs) and anti-thyroglobulin antibodies (TGAs) are the TAA most commonly detected. It has been stated that in a normal population, 6.8-7.0% of people will have detectable circulating TAAs, but in patients presenting with thyroid disease of any form, the prevalence will be much higher.^1,2,6 The literature suggests that TGAs are of less relevance than TMAs in the detection of thyroid disease.^1,3,6

Studies have related the presence or absence of detectable TAAs to the risk of subsequent hypothyroidism after bilateral subtotal thyroidectomy for thyrotoxicosis. These have shown no association, hypofunction being totally dependent on subsequent remnant size.^7,8

Although the possible occurrence of hypothyroidism following hemithyroidectomy is established,^9-11 the use of preoperative TAAs, as a predictor of hypothyroidism, has not been adequately described. The present study looks at the risk of hypothyroidism after hemithyroidectomy for non-toxic nodular goitre, and questions whether removal of half the gland in a patient with detectable TAAs and/or lymphocytic infiltration in the glandular tissue surrounding the thyroid nodule(s) may precipitate thyroid hypofunction.

During the 5-year period from January 1992 to December 1996, 261 patients were operated upon by a variety of forms of unilateral or bilateral thyroidectomy. Of these, 230 were female and 31 were male (F:M = 7.4:1), with an average age of 46.5 years (range 14-90). Two hundred and twenty of these patients had nodular goitre, and were clinically and biochemically euthyroid. One hundred and fifty-eight had a hemithyroidectomy performed for solitary nodular, or predominantly unilateral multinodular, goitre. These 158 patients formed the study group.

Pre-operatively, all patients had free thyroxine (FT4), thyroid stimulating hormone (TSH), TMA and TGA levels measured.

TSH (normal range 0.15-3.5 mU/l) and FT4 (normal range 10-27 pmol/l) were measured by enhanced luminescence techniques of the commercial Amerlite TSH-30 ultrasensitive and MAB FT4 assay kits, respectively (Kodak clinical Diagnostics Ltd.). TMA and TGA levels were detected using the Serodia^®-AMC and ATG kits, respectively, by haemagglutination of gelatin particle carriers. TMA titres were considered weakly positive at titres of between 1/400 and 1/1600, and strongly positive >1/6400.ł

Specimens for histology which had been fixed in formalin, embedded in paraffin, and stained with haematoxilin and eosin were re-examined by one pathologist. In addition to the histology of the thyroid nodule(s), the presence or absence of diffuse lymphocytic infiltration into the surrounding thyroid tissue was determined. Scattered lymphocytes surrounding a nodule, indicating reactive hyperplasia, were not considered significant.

For the purposes of this study, all patients were followed up for a period of a minimum of 2 years, being seen at intervals post-operatively of 4 weeks, 3 months, 6 months and 1 year. At each visit, TSH and FT4 levels were measured. Subclinical hypothyroidism was defined for the purposes of this study as an elevated TSH but a normal FT4 level. Patients who had persisting elevation of TSH (>3.5 mU/l) over at least a 3 month period, or those who had a low FT4 (<10 pmol/l) with an elevated TSH, were treated with thyroxine.

Statistical analysis was performed using the chi-squared test (c2), with P<0.05 considered as significant.

The overall prevalence of TAAs amongst all 261 patients with thyroid disease seen in the unit was 34.1%. Of the 220 patients with nodular goitre, 23.1% had detectable TAAs, the incidence being higher in differentiated carcinomas (28.6%) than in solitary benign (22.7%) or cystic (4.8%) nodules.

Figure 1 demonstrates the distribution of TAAs according to the histopathology of the goitre in 259 of the 261 patients. Medullary carcinoma and metastatic carcinoma have been excluded because of the small numbers of patients. Of those cases with benign solid nodular goitre (i.e. follicular adenoma and multinodular goitre), 22.7% had TAAs, and 28.1% of the malignant solid tumours had circulating levels of TAAs.

Figure 1: Distribution of TAAs according to pathology amongst 259 patients undergoing all forms of thyroidectomy. This excludes two patients, one with medullary carcinoma, and one with metastatic carcinoma. Key: Fol. ad. - Follicular adenoma; Fol. ca. - Follicular carcinoma; Pap. ca. - Papillary carcinoma; MNG - Multinodular goitre; Thyroiditis - Lymphocytic thyroiditis; Graves’ - Graves’ disease; figures in parentheses represent per cent of positive TAAs in each group of pathological disorders

Of the 89 (34.1%) patients with TMAs, 26 had elevated levels of TGAs. None of the patients who were TGA-positive had a negative TMA assay. There was no significant difference between sex with regard to TAA status. Forty-one (25.9%) of the 158 patients undergoing hemithyroidectomy had elevated levels of TAAs.

TSH levels amongst patients undergoing hemithyroidectomy peaked at 3-6 months post-operatively. All 158 patients were followed-up for a minimum of 2 years. Thirty-eight patients (24.1%) became hypothyroid, as defined above. Twenty-two of the 41 TAA-positive patients became hypothyroid within 2 years (53.7%) (Table 1a). Thirty-one patients had high levels of TMAs, of whom 20 (64.5%) required post-operative thyroxine replacement (Table 2).

Table 1: (a) post-operative thyroxine requirements and (b) pathological and immunological findings amongst all 158 hemithyroidectomy cases

a)	Lymphocyte infiltration		Total	TAA
a)	Yes	No	Total	Yes	No
Thyroxine required:
Yes	22	16	38	22	16
No	21	99	120	19	101
Total	43	115	158	41	117
P-value	<0.001			<0.001

b)	Lymphocyte infiltration		Total
b)	Yes	No	Total
TAA
Yes	33	8	41
No	10	107	117
Total	43	115	158
P-value	<0.001

Table 2: Strength of TMA and thyroxine requirements amongst TAA-positive hemithyroidectomy patients

a)	TMA		Total
a)	Strong	Weak	Total
Thyroxine required:
Yes	20	2	22
No	11	8	19
Total	31	10	41
P-value	<0.02

Amongst the 158 hemithyroidectomy patients, 33 of the 41 TAA-positive patients (80.5%) had lymphocytic infiltration into the gland surrounding the thyroid nodule(s) (Table 1b). Ten further patients had lymphocytic infiltration into the gland but were TAA-negative. Table 3a demonstrates that only 4 of these 10 patients lacking TAAs (40%) became hypothyroid post-operatively.

Forty-three patients had lymphocytic infiltration in the surrounding gland, and 22 (51.1%) of these required thyroxine replacement during the follow-up period (Table 1a). Of the 115 who had no significant lymphocytic infiltration, only 16 (13.9%) required thyroxine replacement (Table 1a). In none of these patients did the histology of the nodule(s) relate to the occurrence of hypothyroidism post-opera-tively. However, of these 16 patients with no lymphocytic infiltration who became hypothyroid, 4 (25%) were TAA-positive (Table 3b).

Table 3: Post-operative thyroxine requirements in hemithyroidectomy patients, as predicted by lymphocytic infiltration or TAA as independent variables.

	Lymphocyte infiltration (TAA negative)		Total no lymphocytic infiltration		TAA	Total
	Yes	No	Yes	No
Thyroxine required:
Yes	4	12	16		12	16
No	6	95	101		98	99
Total	10	107	117		110	115
P-value	<0.001				<0.001

The simultaneous use of the presence of TAAs in the circulation and lymphocytic infiltration of the thyroid gland as a predictor of post-operative hypothyroidism was of no greater statistical relevance when compared with the independent use of these parameters.

The prevalence of TAAs in the current study (34.1%) is greater than that documented for a healthy population (6.8%2-7.0%).⁶ The present finding of TAAs in 28.6% of malignant tumours falls within the range of that described in the literature (23.0%12-40.0%).¹³ Papillary, follicular and medullary carcinomas have been associated with TAAs in 23.9%, 25.0% and 4.1% of cases, respectively.¹²

Consistent with the results published,^1,5,6,14 this study has shown the limited value of TGAs as indicators of auto-immune thyroid disease. TGAs were never detectable in this series other than in association with TMAs. This study, therefore, used TMA estimations throughout, in correlation with histopathology and clinical follow-up. At the time of writing, the detection of TMAs by the immunological laboratory of the Blood Transfusion Service in Edinburgh was being superseded by the detection of thyroid peroxidase (TPO) auto-antibodies, due to the increased sensitivity of this test. It is known that a close intracellular relationship exists between TPO and thyroid microsomal antigen,¹⁵ and it is now accepted that both are in fact the same entity.¹⁶

In the current series, 24.1% of patients undergoing hemithyroidectomy for a variety of non-toxic nodular pathologies became hypothyroid within 2 years of surgery. Table 1a highlights the significance of the presence of circulating TAAs as a predictor of post-operative hypothyroidism in these patients (P<0.001). The stronger the titre of TMAs, the more significant the association (P<0.02) (Table 2).

In a study of patients undergoing partial thyroidectomy for primary thyrotoxicosis or nodular non-toxic goitre, Griffiths et al (1974) demonstrated that post-operative hypothyroidism in the two groups (11.0% and 6.5%, respectively) was not related to the presence of lymphocytic infiltration, but to the size of the thyroid remnant.7 Other studies looking at Graves’ disease have shown different results. Some studies have documented that lymphoid follicles with germinal centres are more predictive of post-operative hypothyroidism^5,14,17 than lymphocytic infiltration alone^5,8,14,17. However, Irvine et al (1962) have indicated that lymphocytic infiltration per se is a determinant of hypothyroidism, and this is more likely, the greater the extent of the infiltrate.¹⁸

Studies of patients with Graves’ disease undergoing subtotal thyroidectomy have shown that TAAs are associated with hypothyroidism.^8,14,17 Chou et al (1999) concluded that increased TMA levels were the only independent factor causing post-operative hypothyroidism (8%) and even the presence of TGAs was known to result in hypothyroidism.^8,18 Due to the extensive overlapping between TMA levels amongst the hypothyroid and non-hypothyroid patients, it is advocated that patients with strongly positive TMA levels should be left with a larger thyroid remnant.

These aforementioned series contrast with the findings of studies looking at non-toxic nodular goitre. Bang et al (1985) audited 91 patients with non-toxic nodular goitre over 24 post-operative months.¹⁰ Forty-two patients underwent sub-total or total hemithyroidectomy, and 49 underwent bilateral subtotal thyroidectomy. Lymphocytic infiltration was graded quantitatively according to the number of aggregations per slide. The overall incidence of hypothyroidism was 9%. The study suggested that lymphocytic infiltration was an important factor in the development of post-operative hypothyroidism, but that TAA levels were not. It is of note that none of the patients who were TAA-positive and had lymphocytic infiltration became hypothyroid within the 24 month post-operative period. However, that study included a number of unilateral total and subtotal, and bilateral subtotal thyroidectomy patients. It is difficult, therefore, to relate these findings to the results of the present study.

In a group of 177 patients undergoing hemithyroidectomy for non-toxic goitre described by Berglund et al (1991), the incidence of post-operative hypothyroidism was 7.4%.¹¹ In this, lymphocytic infiltration was graded from sparse and irregularly-distributed foci of lymphocytes to diffuse infiltrates of lymphocytes with germinal centres and oxyphil metaplasia. Patients with greater infiltration were more likely to be rendered hypothyroid (33.3%).

Therefore, these studies demonstrate that lymphocytic infiltration is not always the most sensitive predictor of hypothyroidism in patients operated on for thyrotoxicosis.^7,8,14,17,18 It is more reliable, however, for non-toxic nodular goitre, as exemplified by the present series.^10,11

In the current study, the extent of lymphocytic infiltration was not graded. Lymphocytic infiltration was only felt to be of significance if diffuse in the surrounding thyroid. Therefore, those specimens with sparse and irregularly-distributed foci of lymphocytes were not regarded as arising from a background of lymphocytic thyroiditis. In the present series, lymphocytic infiltration was graded as present or absent.

The close association between detectable TAAs and histological lymphocytic infiltration into the surrounding thyroid gland mainly accounts for the high level of hypothyroidism developing in such patients. However, Table 3 indicates that lymphocytic infiltrate (P<0.001) and positive TAAs (P<0.001) are independent predictors of post-operative hypothyroidism. The presence of either, therefore, indicates that long-term follow-up will be required.

Sato et al (1995) have shown that over a 10-year period, 27.0% of patients with euthyroid TAA-positive Hashimoto’s thyroiditis will become hypothyroid.¹⁹ If the natural history of TAA-positive thyroiditis is towards a state of hypothyroidism, it is reasonable to expect that removing half the thyroid gland in such patients will hasten the transition to hypothyroidism. The occurrence of hypothyroidism (24.1%) in the present series of hemithyroidectomies is greater than that reported in previous studies (7.4%¹¹-9.0%⁹). In both latter studies, the upper limit for TSH was 8 mU/l, higher than that used in the current study, and this may account for the greater rate of hypothyroidism in this study. It is possible that the group of patients operated upon in this series were at a greater predisposition to hypothyroidism, as part of the natural history of their nodular goitre, than those in other studies. It should also be noted that hemithyroidectomy was total in all patients in this series, whereas other series included patients who had undergone subtotal hemithyroidectomies.^9,11 The high prevalence of TAAs (25.9%) and lymphocytic infiltration (27.2%) in the patients with clinically solitary thyroid nodules suggests that these develop more frequently on a background of thyroiditis.

The timing of the use of thyroxine to correct post-operative hypothyroidism is controversial. This study intentionally makes no attempt to discuss the benefits of and indications for prescribing thyroxine post-operatively, and at what stage it should be prescribed.

A number of conclusions can be drawn from our study. Estimation of TGAs is not essential in the diagnosis of auto-immune thyroid disease. Patients with nodular goitre have a higher prevalence of TAAs than the average population. TAA-positive patients with nodular goitre are more likely to become hypothyroid following unilateral total thyroid lobectomy than TAA-negative patients (P<0.001), this association being related to the level of TMAs documented (P<0.02). Histological assessment showing diffuse lymphocytic infiltration of the thyroid tissues surrounding a thyroid nodule is associated with elevated TAAs (P<0.001) and increased risk of post-hemithyroidectomy hypothyroidism (P<0.001). It is essential, therefore, that patients with nodular goitre who are treated by hemithyroidectomy should have TAA estimations performed preoperatively. In view of the fact that Hashimoto’s thyroiditis associated with elevated TAA levels has a tendency to hypothyroidism, particular attention should be made to ensure long-term follow-up in those who are TAA-positive.^1-3 It is unlikely that thyroid function will deteriorate further in those who are TAA-negative or have no lymphocytic infiltration and this group of patients, therefore, does not require such long-term review. Patients whose histology demonstrates a significant lymphocytic infiltration into the thyroid tissues surrounding a thyroid nodule also require long-term follow-up after hemithyroidectomy (P<0.001). The combined use of lymphocytic infiltration and TAA-positivity is no more a sensitive predictor of hypothyroidism than either factor alone.

We wish to thank Dr K. McLaren, Consultant Pathologist, University of Edinburgh Medical School for her helpful criticism of the manuscript, and for review of histological specimens; and Mrs. J. Sawers and the Blood Transfusion Service, Edinburgh for providing pre-operative TAA data on patients’ serum samples. This paper was presented by Malcolm A. Buchanan at the British Association of Endocrine Surgeons, May 1999.

Correspondence: Mr D. Lee, Consultant Surgeon, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh EH3 9YW, U.K.

Thyroid auto-antibodies, lymphocytic infiltration and the development of postoperative hypothyroidism following hemithyroidectomy for non-toxic nodular goitre