The Value of Color Flow Doppler Ultrasonography of the Superior Thyroid Artery in the Surgical Management of Graves Disease FREE

THE DEGREE of vascularization was widely applied in various physiological and pathologic conditions such as diabetic retinopathy, tumor formation, and autoimmune diseases.^1- 2 Although the histopathologic microvessel density (MD) using an immunohistochemical technique is the current "gold standard" to characterize vascularization,² it may not be the ideal tool for clinical purposes because it needs to be performed on biopsy specimens. Noninvasive assessment of vascularity is possible in vivo using color flow Doppler ultrasonography (CFDS), dynamic contrast-enhanced magnetic resonance imaging, and positron emission tomography.^3- 6 These methods may be preferable to histological assay because they are noninvasive, survey the entire organ, reflect both anatomical and physiologic characteristics, and may be useful to monitor the treatment.

With CFDS in vivo vascularity can be assessed rapidly and noninvasively. The technique has been successfully applied in distinguishing between the vascularity in benign and malignant tumors⁷ and in correlating the biological properties of human cancer.⁸ Various thyroid disorders, including Graves disease, can modify thyroid blood flow, which was conspicuously correlated with levels of free triiodothyronine⁹ and iodine uptake.¹⁰ Color flow Doppler ultrasonography has found higher intrathyroidal blood flow^11- 14 and in creased peak systolic velocity¹¹ and blood velocity¹⁴ in patients with Graves disease. The aims of the present study were to examine the factors affecting blood flow through the hypervascular thyroid glands of patients with Graves disease to determine the usefulness of CFDS in determining blood flow in diseased thyroid glands, and to evaluate the value of this assessment in the management of patients with Graves disease.

A total of 52 patients (3 men and 49 women; age range, 19-43 years; mean age, 30.5 years) with Graves disease were enrolled in this prospective study between June 1, 1999, and December 31, 2000. In all cases, the diagnosis was based on the usual clinical criteria and laboratory test results of the following: free thyroxine level, triiodothyronine level, thyrotropin level, thyrotropin–receptor autoantibody level, and anti–thyroid peroxidase antibody level.

The standard treatment protocol was as follows: all patients received methimazole and propranolol hydrochloride (Inderal) preoperatively. Patients received surgical intervention when euthyroidism had been achieved. Color flow Doppler ultrasonography (SSA-340A; Toshiba, Tochigi-ken, Japan) using a 10-MHz linear transducer was done for all patients in the morning of the operative day. Fifty-two age- and sex-matched healthy medical staff volunteers were also measured using CFDS; they served as healthy control subjects. The superior thyroid artery on each side of the neck was located first, then measurements of the time-averaged blood velocity, vessel diameter, and the angle between the ultrasound beam and vessel axis were taken and entered into the computer that then calculated the blood flow rate of the superior thyroid artery in milliliters per minute.

The operative procedure (total thyroidectomy of 1 lobe together with subtotal thyroidectomy of the other) was performed by the same surgeon (S.-M.H.). The weight of the resected gland was measured during operation. Three grams of thyroid remnant, estimated by using 3 g of weighed tissue from the contralateral resected thyroid, was preserved in situ. The total weight of the thyroid gland was the weight of resected gland plus 3 g. The volume of blood loss (expressed in milliliters), measured as the amount of blood in the suction bottle and absorbed by the gauze, was recorded for each patient.

Thyroid gland specimens were examined by 1 of us (N.-H.C.) without knowledge of the CFDS findings. The histopathologic changes were classified into 3 types of microscopic pattern. Those cases having apparent diffuse microfollicular hyperplastic follicles were categorized histopathologically as the compact pattern (Figure 1A); the colloid pattern (Figure 1B) was composed predominantly of cystically dilated follicles with flattened epithelial lining; and the mixed pattern (Figure 1C) was composed of the compact and colloid patterns.

The MD was assessed by immunohistochemical expression of factor VIII in thyroid biopsy specimens because it correlates well with patient outcome and it is comparable with immunohistochemical expression of the CD34 molecule.¹⁵ Immunohistochemical staining was performed as described in detail by Yamamoto et al.¹⁵ Briefly, monoclonal anti–factor VIII antibody (diluted 1:100; Calbiochem, Cambridge, Mass) was incubated with thyroid gland tissue. Sections were washed for 5 minutes with a buffered-saline solution followed by incubation with 3% hydrogen peroxide at room temperature. Then the sections were covered with 3% normal horse serum for 15 minutes. Next, the second antibody was incubated with these sections at 4°C overnight. A commercially available kit (LSAB2; Dako Co, Carpinteria, Calif) was used to detect the immune complex, and the activity was visualized using the aminoethyl carbazole substrate solution kit (Zymed Laboratory, Inc, San Francisco, Calif) according to the manufacturer's directions. Finally, sections were counterstained with hematoxylin–eosin. Negative control sections were incubated with nonimmune rabbit IgG instead of primary antibody. The number of microvessels highlighted by factor VIII expression was counted in a total of 20 high-power fields (or entire slide) for each patient. The density of positive staining (vessels per high-power field [×200]) was determined from the total vessel numbers divided by the number of high-power fields (Figure 2).

Results were analyzed by 1-way analysis of variance (ANOVA), 2-sample t test ANOVA, regression analysis, and the Pearson product moment correlation. All values are expressed as mean (SD).

The blood flow rate of the superior thyroid artery of patients with Graves disease is significantly higher than that of the healthy controls (97.5 [52.7] mL/min vs 20.1 [5.7] mL/min, P<.001). In Graves disease, blood flow rate was positively correlated with thyroid gland weight (Figure 3, P = .03) or MD (Figure 4, P<.01) using the Pearson product moment correlation. Regression analysis was performed on the relation expressed as follows:

Histolopathogic change and vascularization were examined only in the surgical specimens from patients with Graves disease; wide variation was noted in each case. Microvessel density showed a positive correlation with histopathologic patterns (P<.01) by ANOVA. The relationship of histopathologic pattern distribution to MD and blood flow are shown in Figure 5. Diffuse hyperplastic thyroid tissue (compact pattern) had significantly higher vascular density and blood flow than did either of the other 2 types.

Interestingly, 10 patients having a blood loss exceeding 200 mL during thyroidectomy showed a higher preoperative blood flow rate (Figure 6) and vascular density (Figure 7) (P<.01) than those with less blood loss (n = 42). Of the 8 patients whose blood flow was more than 0.15 L/min, 6 (75%) had a blood loss exceeding 200 mL during surgery and 1 patient who had a 1000-mL blood loss needed a blood transfusion (Figure 8).

Graves disease is a genetically determined autoimmune disorder characterized by hyperthyroidism. The laboratory diagnosis is based on the finding of high levels of serum thyroid hormones and undetectable serum thyrotropin associated with circulating antibodies to thyrotropin receptor and thyroperoxidase. Untreated patients have a hypervascular gland accompanied by bruit and thrill. Consistent with this notion are prior reports^11- 14 in which higher intrathyroidal blood flow and increased peak systolic velocity were detected in association with Graves disease using CFDS. The present study of the blood flow of the superior thyroid artery and intrathyroid MD further supports this concept.

Previous studies reported that thyroid artery blood flow correlated remarkably with the levels of free triiodothyronine,⁹ whereas others did not find a correlation between CFDS findings and serum thyroid hormone levels.¹⁴ We found that CFDS-monitored blood flow through the superior thyroid artery was associated with intrathyroid MD, as Castagnone et al¹² previously reported. Of particular interest was the significant correlation between blood flow rate and histopathologic microscopic patterns. The blood flow rates on CFDS were higher in those with persistent hyperplastic histopathologic conditions than in patients with changes in thyroid colloid. This finding is similar to observations noted by Nagura et al² and Morosini et al.¹⁶ Nagura et al² and Iitaka et al³ found that increasing vascular density positively correlated with vascular endothelial growth factor expression in hyperplastic follicular cells and serum vascular endothelial growth factor levels, respectively. Those findings may support the association of angiogenesis with disease activity of Graves disease.

Both the histopathologic features of the thyroid gland and vascular density differed considerably even though euthyroidism had been achieved. The findings seem to imply that a mechanism, other than the protirelin-thyrotropin-thyroxine–axis, is functioning during the development of follicular hyperplasia and increased vascularity of Graves disease. Although Castagnone et al¹² found that those with higher blood flow (in Graves disease remission due to antithyroid drug therapy) had a higher rate of recurrence, hyperthyroidism has not recurred in our patients so far, irrespective of the histopathologic condition of the thyroid gland and blood flow. Nonrecurrence might be explained by the relatively small size of the remnant tissue (3 g).

Both surgical dissection and hemostasis are 2 major concerns for surgeons during thyroidectomy. Graves disease is characterized histologically by thyroid friability and vascularization. We found that patients with a higher preoperative blood flow (>150 mL/min) are at risk of intraoperative blood loss exceeding 200 mL. The result seems rational since the bleeding rate from cut vessels can be expected to increase when blood flow is higher. The oozing increases when the MD is higher. Although radioiodine therapy is the most common treatment for Graves disease and surgery has become uncommon in the United States, recent articles^17- 18 have emphasized the advantages of surgery; moreover, surgery still plays an important role in Asia and Europe.^19- 20 If surgery is indicated for treatment of Graves disease, our findings suggest that preoperative CFDS may be useful in the preoperative assessment of intraoperative bleeding and in identifying patients who should undergo preoperative autodonation of blood or be cross-matched for banked blood in this era of limited elective blood supply.

Chang et al²¹ directly demonstrated that blood flow in patients with Graves disease was reduced within the superior thyroid arteries after administering Lugol potassium iodide solution. Ansaldo et al²² also revealed a decrease in arterial resistance after administering Lugol potassium iodide solution. Because of the bleeding risks associated with cutting and dissection for such a hypervascular disorder, the routine preoperative use of Lugol potassium iodine solution is recommended for all patients with Graves disease; furthermore, the use of more sophisticated hemostastic instruments, such as bipolar or ultrasonic cauterization,^23- 24 is also a rational approach, especially for those patients with higher superior thyroid artery blood flows.

We demonstrate that the blood flow through the superior thyroid artery positively correlates with intrathyroid MD, glandular weight, and the histopathologic microscopic pattern. Patients with diffuse hyperplastic glands or who have more blood loss during thyroidectomy have significantly higher blood flow and higher MD. Preoperative CFDS may aid in identifying those patients who are liable to bleed intraoperatively during thyroidectomy.

Corresponding author and reprints: Shih-Ming Huang, MD, Department of Endocrine Surgery, Buddhist Tzu Chi General Hospital and Tzu Chi University, 707, Chung Yan Rd, Section 3, Hualien, Taiwan (e-mail: smhuang@mail.tcu.edu.tw).

Accepted for publication October 5, 2002.