Predictive factors for intraoperative excessive bleeding in Graves’ disease

Summary

Background

In Graves’ disease, because a thyroid tends to have extreme vascularity, the amount of intraoperative blood loss (AIOBL) becomes significant in some cases. We sought to elucidate the predictive factors of the AIOBL.

Methods

A total of 197 patients underwent thyroidectomy for Graves’ disease between 2002 and 2012. We evaluated clinical factors that would be potentially related to AIOBL retrospectively.

Results

The median period between disease onset and surgery was 16 months (range: 1–480 months). Conventional surgery was performed in 125 patients, whereas video-assisted surgery was performed in 72 patients. Subtotal and near-total/total thyroidectomies were performed in 137 patients and 60 patients, respectively. The median weight of the thyroid was 45 g (range: 7.3–480.0 g). Univariate analysis revealed that the strongest correlation of AIOBL was noted with the weight of thyroid (p < 0.001). Additionally, AIOBL was correlated positively with the period between disease onset and surgery (p < 0.001) and negatively with preoperative free T4 (p < 0.01). Multivariate analysis showed that only the weight of the thyroid was independently correlated with AIOBL (p < 0.001). Four patients (2.0%) needed blood transfusion, including two requiring autotransfusion, whose thyroids were all weighing in excess of 200 g. The amount of drainage during the initial 6 hours and days until drain removal was correlated positively with AIOBL (p < 0.001, each). Occurrences of postoperative complications, such as recurrent laryngeal nerve palsy or hypoparathyroidism, and postoperative hospital stay were not correlated with AIOBL.

Conclusion

A huge goiter presented as a predictive factor for excessive bleeding during surgery for Graves’ disease, and preparation for blood transfusion should be considered in cases where thyroids weigh more than 200 g.

Keywords

Graves’ disease;intraoperative blood loss;thyroidectomy

1. Introduction

Graves’ disease (GD) is an autoimmune disease, in which thyroid-receptor antibodies combine with the thyrotropin receptor, promoting the production of thyroid hormones in the follicular cells through stimulation of cyclic adenosine monophosphate.¹ GD is the most common cause of hyperthyroidism. Acceptable therapeutic options are antithyroid drugs (ATDs), radioiodine (RI) therapy, and surgery.^{1 ;  2} The prevalence of therapy varies in different parts of world. ATDs are the first-line therapy in most of the Japanese patients, whereas RI is preferred in the United States.¹ Although the total number of patients in Japan who received RI therapy was much fewer than that in Europe or the United States,³ the application of RI therapy has extended due to the new regulation that allows its administration in outpatient clinics.⁴ Nevertheless, suspicion of malignancy, unwillingness to receive RI therapy, and pregnancy should still be good indications for surgery. Although thyroidectomy has been regarded as a prompt and assured therapy for GD, it has some drawbacks, such as operative scar, need for hospitalization, and complications related to operation, including recurrent laryngeal nerve (RLN) paralysis, hypoparathyroidism, and intraoperative bleeding.^{1 ;  2} Especially in GD, the thyroid gland tends to have hypervascularity, and hemostasis is often difficult.^{5 ;  6} In this study, we evaluated the factors that are relevant to the amount of intraoperative blood loss (AIOBL), and how AIOBL affects postoperative courses retrospectively.

2. Methods

We reviewed the clinical and laboratory data of 197 consecutive patients with GD, who underwent thyroidectomy between April 2002 and March 2012 at our department. Patients who had undergone neck surgery previously were excluded. GD was defined by the presence of thyrotoxicosis, a diffuse goiter, and a serum thyroid-stimulating hormone receptor antibody or thyroid-stimulating antibody. After the operation, the diagnosis was confirmed histologically in all patients. In preparation for surgery, all patients were given potassium iodide for a median period of 14 days. Operations were performed with conventional open surgery or video-assisted neck surgery. The indications and technique of video-assisted neck surgery were described previously.⁷ Subtotal thyroidectomy was defined as leaving between 2 g and 6 g thyroid tissue and near-total thyroidectomy was defined as leaving < 2 g thyroid tissue.

We placed closed suction drainage tubes prior to wound closure in all of the patients except one and removed them 1–3 days later. RLN injury was recorded if voice change was recognized. Calcium supplements and vitamin D analogs were given to those with symptomatic hypocalcemia. Patients who could discontinue the supplements after achieving normocalcemia within 12 months after surgery were identified as those with transient hypocalcemia, whereas patients who had to continue it for more than 12 months and continued to have a below-normal serum intact parathyroid hormone level were categorized as those with permanent hypoparathyroidism.

Clinical variables were assessed in relation to AIOBL. Univariate analysis was performed by single regression analysis to test the correlation between continuous variables and AIOBL, and Student t test was used to compare between groups. A p value of <0.05 was considered statistically significant. For multivariate analysis, the factors identified as being associated with p ≤ 0.10 were entered into a stepwise regression analysis to determine the independent risk factors for AIOBL. Statistical analysis was performed with StatMate III for Macintosh (ATMS Co., Ltd, Tokyo, Japan).

3. Results

3.1. Background of patients and operative variables

Background of patients and operative variables are shown in Table 1. The majority were female (76.1%), with a median age of 34 years. The median period between the onset of the disease and operation was 16 months. The most frequent indication for operation was intolerance to ATDs (41.2%), followed by patient preference (25.4%) and uncontrollable disease (25.4%). Thirty-seven patients (18.8%) needed preoperative adrenocortical hormone administration due to an excessive value of serum-free T4 even after potassium iodide intake. Median intraoperative blood loss was 100 mL (range: 10–1390 mL), and four patients required blood transfusion. Thyroidectomy was performed by open surgery in 125 patients and video-assisted neck surgery in 72 patients. In terms of the extent of resection, 137 subtotal thyroidectomies and 60 near-total/total thyroidectomies were performed. The median weight of thyroid was 45 g (range: 7–480 g). Transient hypocalcemia was seen in 19.8% of the patients, but permanent hypoparathyroidism in only 1.0% of patients. RLN palsy was observed in 3.0% of patients, all of which were temporary. During the study period, we have not experienced hematoma requiring surgical evacuation. Patients were discharged after their operation within a median period of 4 days (range: 2–22 days).

Table 1. Background of patients and operative variables.

Background of patients
Age (y)	34.0 (11–74)
Sex, male:female	47:150
Period between disease onset and surgery (mo)	16 (1–480)
Indication for surgery
Intolerance to antithyroid drug	81 (41.2)
Patient preference	50 (25.4)
Uncontrollable disease	50 (25.4)
Large goiter with pressure symptoms	41 (20.8)
Poor adherence	20 (10.2)
Graves' ophthalmopathy	10 (5.1)
Planned pregnancy	6 (3.0)
Suspicion of malignancy	2 (1.0)
Thyrotoxic crisis	2 (1.0)
Young age	1 (0.5)
Failure of RI therapy	1 (0.5)
Preoperative TSH at initial consultation (ng/mL)	0.005 (0.0–77.6)
Preoperative free T4 at initial consultation (ng/mL)	2.6 (0.2–21.3)
Preoperative steroid	37 (18.8)
Preoperative β-blocker	79 (40.1)
Variables of surgery
Duration of surgery (min)	186 (69–380)
Amount of blood loss (mL)	100 (10–1390)
Blood transfusion	4 (2.0)
Operative methods
Open surgery:VANS	125:72
Subtotal:near-total/total thyroidectomy	137:60
Weight of thyroid (g)	45 (7–480)
Surgical complications
Transient hypocalcemia	40 (19.8)
Permanent hypoparathyroidism	2 (1.0)
Transient RLN palsy	6 (3.0)
Permanent RLN palsy	0 (0.0)
Amount of drainage during initial 6 h (mL)	41 (6–114)
Removal of drain (d)	1 (1–3)
Postoperative hospital stay (d)	4 (2–22)

Data are presented as n (%) or median (range), unless otherwise indicated.

RI = radioiodine; RLN = recurrent laryngeal nerve; TSH = thyroid-stimulating hormone; VANS = video-assisted neck surgery.

3.2. Factors affecting AIOBL

Table 2 shows the correlations between clinical variables and AIOBL, as shown by univariate analysis. The strongest correlation of AIOBL was found with the weight of thyroid, with a correlation coefficient of 0.72 (p < 0.0001). Additionally, AIOBL was correlated positively with the period between disease onset and surgery (correlation coefficient 0.36, p < 0.001), and negatively with preoperative free T4 at the initial consultation (correlation coefficient –0.22, p < 0.01). Multivariate analysis revealed that only the weight of the thyroid was correlated with AIOBL (p < 0.0001; Table 3). Out of four patients who needed intraoperative blood transfusion (Table 4), two underwent autotransfusion with preoperatively collected blood.

Table 2. Correlation between clinical variables and intraoperative blood loss (univariate analysis).

Continuous variables (single regression analysis)	Correlation coefficient	p
Age (y)	0.08	0.28
BMI	0.05	0.47
Period between disease onset and surgery	0.35	<0.001
Free T4 at initial consultation	0.22	<0.05
Weight of thyroid	0.72	<0.0001

Categorical variables (Student t test)	Amount of bleeding (mL)	p
Sex
Male	120
Female	107	0.55
Operative methods
Open	100
VANS	110	0.1
Type of surgical resection
Subtotal	110
Near total/total thyroidectomy	86	0.51
Weight of thyroid
<100 g∗	90
>100 to ≤200 g∗∗	257
>200 g ∗∗∗	560

∗versus ∗∗, p < 0.05.

∗ versus ∗∗∗ and ∗∗ versus ∗∗∗, p < 0.001.

BMI = body mass index; VANS = video-assisted neck surgery.

Table 3. Correlation between clinical variables and intraoperative blood loss (multivariate analysis).

Multiple regression analysis	95% confidence interval	p
Period between disease onset and surgery	−0.52–0.25	0.48
Free T4 at initial consultation	−12.61–3.8	0.29
Weight of thyroid	1.79–2.48	<0.0001

Table 4. Cases with intraoperative blood transfusion.

Case	Sex/age (y)	Extent of resection	Weight of thyroid (g)	Amount of bleeding (mL)	Type of transfusion
1	Female/30	Total	245	550	Auto
2	Male/48	Total	480	1390	Allogeneic
3	Male/23	Subtotal	369	480	Auto
4	Female/53	Total	324	1250	Allogeneic

3.3. Impact of intraoperative blood loss on postoperative courses

Although AIOBL was correlated positively with the amount of drainage (correlation coefficient 0.47, p < 0.0001) and the duration of drain retained (correlation coefficient 0.25, p < 0.001), no correlation was seen between AIOBL and postoperative hospital stay ( Table 5). Furthermore, the occurrence of postoperative morbidity was not correlated with AIOBL.

Table 5. Impact of blood loss on perioperative courses.

Continuous variables (single regression analysis)	Correlation coefficient	p
Amount of drainage during initial 6 h	0.47	<0.0001
Duration of drain retained (d)	0.25	<0.001
Postoperative hospital stay (d)	0.05	0.48

Categorical variables (Student t test)	Amount of bleeding (mL)	p
Surgical complications
Transient hypocalcemia
Yes	110
No	100	0.67
Transient RLN palsy
Yes	80
No	100	0.15

RLN = recurrent laryngeal nerve.

4. Discussion

Thyroidectomy for GD has been one of the acceptable therapeutic options, although its prevalence differs from country to country.¹ RI therapy is regarded as a safe and effective therapeutic modality; the majority of patients in the United States are treated with RI therapy,^{1 ;  8} which has become a major therapeutic option in Japan as well. Surgery, however, still plays an important role, and thyroidectomy has certain advantages, such as a rapid therapeutic effect, a lower recurrence rate than RI therapy or ATDs,^{8; 9 ;  10} and especially no risk of recurrence after the removal of the whole gland.⁹ Moreover, when ATDs are intolerable, surgery is the only possible therapeutic modality in patients who are unwilling to receive RI therapy, have Graves’ ophthalmopathy, and are pregnant.^{1 ;  2}

In the present study, a huge goiter was the only clinical variable that had a positive correlation with AIOBL, as revealed by multivariate analysis. The thyroid gland in a GD patient is rich of vascularity and has dilated vessels around it, and is therefore prone to bleed from the surface or surrounding vessels.⁵ Additionally, hemostasis can be difficult due to the huge size of the goiter, difficulty in mobilizing the thyroid, reaching vasculatures, and subsequently cutting off the inflow of blood.

We experienced two cases of permanent hypoparathyroidism (1.0%) and no permanent RLN palsy postoperatively, whose rates were comparable to those of previous reports.^{2; 9 ;  11} Because the number of patients with these permanent complications was very small to address statistical evaluation, we evaluated transient complications in this study. Among benign goiter surgeries, GD itself was shown as an independent risk factor for either postoperative transient hypoparathyroidism¹² or RLN palsy.¹³ In our study, AIOBL was not related to the complications, and that might be the reason why no correlation was found between AIOBL and postoperative stay.

Continued growth of a Graves’ goiter in a patient on ATDs and increasing the difficulty of a delayed thyroidectomy were reported previously.² In our analysis, the disease period was correlated positively with the weight of the thyroid and inversely with free T4 levels, according to univariate analysis (data not shown). This implied that the indication for thyroidectomy changed with the course of the disease; for example, although patients with a short term of disease tended to undergo thyroidectomy due to functional issues, those suffering from the disease for a long period of time underwent surgery due to other reasons, such as huge goiter and Graves’ ophthalmopathy, even when function was controlled by ATDs. Several randomized trials showed that administration of ATDs for more than 18 months did not improve the remission rate.^{14; 15 ;  16} Although Erbil et al¹⁷ showed a negative correlation between ADT treatment duration and AIOBL by multivariate analysis, our result was the opposite, according to univariate analysis, even though most of our patients took ATDs at the time of operation. The reason for this discrepancy was not defined; according to our results, definitive therapy should be considered after a couple of years of ATD therapy, not only from the aspect of remission rate but also the safety on operation.

Four out of 11 cases with thyroids weighing more than 200 g underwent intraoperative blood transfusion (data not shown). We could not evaluate the vascularity, which was reported to affect AIOBL,¹⁷ in all cases because it was a retrospective study. We believe that further prospective examinations may find a notable formula that predicts AIOBL more accurately. Regardless of this limitation, our results recommend that preparation for blood transfusion should be made in GD patients with a huge goiter estimated to weigh more than 200 g.

References

1. 1 J. Yip, B.H. Lang, C.Y. Lo; Changing trend in surgical indication and management for Graves' disease; Am J Surg, 203 (2012), pp. 162–167
2. 2 D.V. Feliciano, J.D. Lyons; Thyroidectomy is optimal treatment for Graves' disease; J Am Coll Surg, 212 (2011), pp. 714–720
3. 3 H.J. Baskin, R.H. Cobin, D.S. Duick, et al.; American Association of Clinical Endocrinologists medical guidelines for clinical practice for the evaluation and treatment of hyperthyroidism and hypothyroidism; Endocr Pract, 8 (2002), pp. 457–469
4. 4 J. Tajiri; Radioiodine treatment in patients with Graves’ disease at outpatient clinic: special reference to safety and short-term outcome; Jpn J Nucl Med, 42 (2005), pp. 115–122 [In Japanese with English abstract]
5. 5 K.C. Welch, C.R. McHenry; Total thyroidectomy: is morbidity higher for Graves' disease than nontoxic goiter?; J Surg Res, 170 (2011), pp. 96–99
6. 6 G.J. Kahaly, S.K. Grebe, M.A. Lupo, N. McDonald, J.A. Sipos; Graves' disease: diagnostic and therapeutic challenges (multimedia activity); Am J Med, 124 (2011), pp. S2–S3
7. 7 S. Maeda, T. Uga, N. Hayashida, K. Ishigaki, J. Furui, T. Kanematsu; Video-assisted subtotal or near-total thyroidectomy for Graves' disease; Br J Surg, 93 (2006), pp. 61–66
8. 8 C.M. Schussler-Fiorenza, C.M. Bruns, H. Chen; The surgical management of Graves' disease; J Surg Res, 133 (2006), pp. 207–214
9. 9 J. Liu, A. Bargren, S. Schaefer, H. Chen, R.S. Sippel; Total thyroidectomy: a safe and effective treatment for Graves' disease; J Surg Res, 168 (2011), pp. 1–4
10. 10 S. Grodski, P. Stalberg, B.G. Robinson, L.W. Delbridge; Surgery versus radioiodine therapy as definitive management for Graves' disease: the role of patient preference; Thyroid, 17 (2007), pp. 157–160
11. 11 F.Y. Chiang, J.C. Lin, C.W. Wu, et al.; Morbidity after total thyroidectomy for benign thyroid disease: comparison of Graves' disease and non-Graves' disease; Kaohsiung J Med Sci, 22 (2006), pp. 554–559
12. 12 O. Thomusch, A. Machens, C. Sekulla, et al.; Multivariate analysis of risk factors for postoperative complications in benign goiter surgery: prospective multicenter study in Germany; World J Surg, 24 (2000), pp. 1335–1341
13. 13 S.N. Karamanakos, K.B. Markou, K. Panagopoulos, et al.; Complications and risk factors related to the extent of surgery in thyroidectomy. Results from 2,043 procedures; Hormones (Athens, Greece), 9 (2010), pp. 318–325
14. 14 H. Allannic, R. Fauchet, J. Orgiazzi, et al.; Antithyroid drugs and Graves' disease: a prospective randomized evaluation of the efficacy of treatment duration; J Clin Endocrinol Metab, 70 (1990), pp. 675–679
15. 15 R.V. Garcia-Mayor, C. Paramo, R. Luna Cano, L.F. Perez Mendez, J.C. Galofre, A. Andrade; Antithyroid drug and Graves' hyperthyroidism. Significance of treatment duration and TRAb determination on lasting remission; J Endocrinol Invest, 15 (1992), pp. 815–820
16. 16 D. Maugendre, A. Gatel, L. Campion, et al.; Antithyroid drugs and Graves' disease—prospective randomized assessment of long-term treatment; Clin Endocrinol (Oxf), 50 (1999), pp. 127–132
17. 17 Y. Erbil, M. Giriş, A. Salmaslioglu, et al.; The effect of anti-thyroid drug treatment duration on thyroid gland microvessel density and intraoperative blood loss in patients with Graves' disease; Surgery, 143 (2008), pp. 216–225