ORIGINAL ARTICLE – ENDOCRINE TUMORS

Is Potassium Iodide Solution Necessary Before TotalThyroidectomy for Graves Disease?

Myrick C. Shinall Jr., MD1, James T. Broome, MD2, Arielle Baker2, and Carmen C. Solorzano, MD2

1Department of Surgery, Vanderbilt University Medical Center, Nashville, TN; 2Division of Surgical Oncology and

Endocrine Surgery, Vanderbilt University Medical Center, Nashville, TN

ABSTRACT

Background. Potassium iodide (KI) has traditionally been

used to reduce gland vascularity and diminish blood loss in

patients undergoing thyroidectomy for Graves disease

(GD). Current American Thyroid Association (ATA)

guidelines (Recommendation 22) call for its routine

administration in GD but avoidance in toxic multinodular

goiter (TMNG).

Methods. A retrospective review (July 2008–May 2012)

of perioperative data was performed on 162 patients

undergoing total thyroidectomy without preoperative KI

and compared to 102 patients with TMNG. Statistical

analysis included Student’s t test, v2 test, and multivariate

linear regression.

Results. Compared to TMNG patients, GD patients had a

lower mean age (42.7 vs. 49.6 years, p \ 0.001) and were

less likely to be obese (37 vs. 54 %, p = 0.047). No

patients were provided KI in preparation. GD patients did

not differ significantly from TMNG patients with respect to

mean estimated blood loss (55.4 vs. 51.5 mL, p = 0.773)

or mean operative time (131.5 vs. 122.6 min, p = 0.084).

GD patients had a lower rate of transient hypocalcemia (31

vs. 49 %, p = 0.004), but the two groups did not statisti-

cally differ in rates of prolonged hypocalcemia, temporary

recurrent laryngeal nerve (RLN) palsy, prolonged RLN

paralysis, or hematoma formation.

Conclusions. Although current ATA recommendations for

the management of GD call for routine use of KI before

thyroidectomy, this large series demonstrates no appre-

ciable detriment to patient outcomes when this goal is not

met.

Total thyroidectomy (TT) is recognized as a safe and

durable treatment for Graves disease (GD). Preoperative

administration of potassium iodide (KI) has become a

widely used practice whose goal is to increase the safety of

TT in GD. Plummer1 first advocated the use of KI for this

patient population in the 1920s, and the practice gained

widespread acceptance in the following decades. Current

American Thyroid Association guidelines recommend the

routine preoperative administration of KI before TT for

GD.2

The rationale for KI administration stems from its effect

in decreasing thyroid gland blood flow. The increased

vascularity of the thyroid gland in GD patients can make

TT difficult, and excess blood loss impairs visualization,

which may put patients at greater risk for complications.

Administration of KI is thought to mitigate the problems

associated with this increased vascularity. However,

improvements in hemostatic technique during surgery may

obviate the need for preoperative KI in the management of

GD.

Toxic multinodular goiter (TMNG) is another common

etiology of hyperthyroidism for which TT is often per-

formed. Without the continuous stimulation of the TSH

receptor and growth of the gland, increased vascularity is

not a significant problem in TMNG, as opposed to GD. KI

administration therefore is not recommended.

We hypothesized that preoperative KI administration

does not significantly reduce blood loss or rates of com-

plications in TT for GD in the hands of experienced

endocrine surgeons. We retrospectively compared a cohort

of GD patients who did not receive KI before TT with a

cohort of TMNG patients to determine whether the GD

Myrick C. Shinall Jr. and James T. Broome contributed equally to this

article, and both should be considered first author.

� Society of Surgical Oncology 2013

First Received: 22 April 2013;

Published Online: 12 July 2013

M. C. Shinall Jr., MD

e-mail: Ricky.shinall@vanderbilt.edu

Ann Surg Oncol (2013) 20:2964–2967

DOI 10.1245/s10434-013-3126-z

patients had worse outcomes than a comparable group of

patients when neither group received KI preoperatively.

MATERIALS AND METHODS

After approval by the Institutional Review Board, a ret-

rospective analysis of the medical records at a large urban

academic hospital was conducted. All patients undergoing

TT with a preoperative diagnosis of GD or TMNG from 1

July 2007, through 30 May 2012, were included. Information

on preoperative medical history, operative characteristics,

and postoperative complications was collected.

The diagnosis of GD or TMNG was generally estab-

lished by the referring endocrinologist utilizing clinical

signs and symptoms, thyroid hormone and thyroid-stimu-

lating hormone assays, radioactive iodine uptake and scan,

and/or measurement of thyrotropin receptor antibodies. GD

was diagnosed on the basis of biochemical evidence of

hyperthyroidism in the presence of thyrotropin receptor

antibodies and a diffuse goiter on examination and imag-

ing. TMNG was diagnosed on the basis of biochemical

evidence of hyperthyroidism in the absence of thyrotropin

receptor antibodies and a nodular goiter on exam and

imaging. Patients were deemed ready for surgery on the

basis of the judgment of the surgeon in consultation with

the referring endocrinologist if they were not clinically

toxic and treatment had been initiated with antithyroidal

medication and/or beta-blockers. Prethyroidectomy KI was

not routinely used. During the study period, three GD

patients received KI preoperatively, and they were exclu-

ded from the analysis. TT was performed at a single center

by one of six endocrine surgeons. Recurrent laryngeal

nerve (RLN) monitoring was used in all cases. Operative

site drains were not used. Patients were discharged the

same day or after overnight observation unless a compli-

cation developed requiring longer inpatient hospitalization.

Serum calcium levels were obtained the next morning, and

patients were placed on standard calcium supplements (1 g

of elemental calcium TID). Calcium levels were rechecked

2 weeks after thyroidectomy. Calcitriol was added when

calcium levels were \7.8 mg/dL or when the patient was

symptomatic. Parathyroid hormone levels were not rou-

tinely checked unless clinically indicated.

The primary outcome was estimated blood loss (EBL)

during TT. EBL was obtained from the operative and

anesthetic records. The study had an 84 % power to detect

a 40-mL difference in mean EBL between the two groups.

Secondary outcomes were operative time, length of stay,

and incidence of complications. Postoperative hypocalce-

mia was defined as either a serum calcium level below the

reference range or complaints of hypocalcemic symptoms

requiring treatment above routine replacement during the

immediate postoperative period. A patient was considered

to have prolonged postoperative hypocalcemia if bio-

chemical evidence of hypocalcemia or need for ongoing

supplemental calcium salts or vitamin D analogs persisted

beyond 6 months after TT. RLN paresis was defined as

subjective complaints of hoarseness at the first postopera-

tive visit (2 weeks). Prolonged RLN palsy was defined as

hoarseness and documented RLN paralysis at laryngoscopy

persisting beyond 6 months. Hematoma was defined as a

hematoma requiring reopening of the cervical incision.

Statistical inference was carried out by t tests with

unequal variances for continuous variables and the v2 test

for categorical or dichotomous variables. Multivariate

analysis utilized multiple linear regression with Stata

(StataCorp LP, College Station, TX).

RESULTS

During the study period, 165 patients underwent TT for

GD, while 102 patients underwent TT for TMNG. Three

GD patients who received KI preoperatively were excluded

from the analysis. Patient characteristics are listed in

Table 1. The two groups differed significantly only with

respect to age and body mass index (BMI). Standard def-

initions for underweight (BMI \18.5 kg/m2), normal

weight (BMI 18.5–24.9 kg/m2), overweight (BMI 25–

29.9 kg/m2), and obesity (BMI C30 kg/m2) were used for

this analysis. When controlling for BMI, there was no

significant difference in thyroid gland weight between the

two groups; therefore, it does not appear that any differ-

ence in gland size mediated by the higher BMI of the

TMNG group contributed to the findings.

The primary outcome, EBL at TT, did not differ sig-

nificantly between the GD patients and the TMNG patients

(Table 2). The two groups also did not differ significantly

with respect to operative time or length of stay. Transient

hypocalcemia was actually more frequent among the

TMNG patients compared to the GD patients (49 vs. 31 %,

p = 0.004). No other short-term or long-term complication

rates differed significantly between the two groups.

To account for any confounding from the differences in

baseline patient characteristics, multiple linear regression

was used to model the difference in EBL between the two

groups. Controlling for obesity and age (the two character-

istics that were significantly different between the two

groups), the TMNG and GD patients differed in EBL by only

3.2 mL, which was not statistically significant (p = 0.831).

DISCUSSION

There were no statistically significant differences in

EBL, operative time, or length of stay between GD patients

Potassium Iodide Solution in Graves Disease 2965

who did not receive KI preoperatively and TMNG patients.

The only complication that significantly differed between

the two groups was transient hypocalcemia. This compli-

cation favored the GD group. Other complication rates in

both patient groups were comparable to each other and to

those rates reported in the literature for TT in these con-

ditions. In these studies the incidence of transient

hypocalcemia ranged from 3 to 40 %, with the incidence of

long-term hypocalcemia ranging from 0.6 to 6 %. Simi-

larly, the incidence of transient and permanent RLN palsy

ranged from 0 to 24, and 0 to 2 %, respectively.3–13

Plummer 1 first advocated perioperative KI administra-

tion to reduce the risk of thyroid storm in the era before

thionamides and beta-blockers, when thyroid storm was a

major source of morbidity for thyroidectomy. Subsequent

investigators noted that KI administration reduced thyroid

gland vascularity on histologic examination.14 Once beta-

blockers and thionamides made KIs thyrostatic properties

obsolete for the prevention of thyroid storm, the reduction

of thyroid vascularity became the primary justification for

KI administration.15 The evidence for the clinical utility of

this practice has been equivocal. Much of the evidence in

favor of KI administration comes from imaging studies

assessing thyroid blood flow, the clinical implications of

which are difficult to determine.16–18

A number of studies have failed to demonstrate clear

benefits of KI administration. An older double-blind ran-

domized controlled trial found no appreciable difference in

blood loss at thyroidectomy or surgeon reports of operative

difficulty whether or not KI was administered.19 Later

retrospective studies have similarly failed to demonstrate

decreased blood loss in TT for GD with preoperative KI

administration.20,21 The current study adds to this literature

by demonstrating that GD patients who do not receive KI

have similar outcomes to TMNG patients who undergo TT

by the same surgeons.

The most compelling evidence in favor of KI adminis-

tration comes from an unblinded, randomized trial by Erbil

et al.22 In this trial 36 patients with GD were assigned

either to preoperative treatment with KI or to no treatment.

The investigators found a statistically significant difference

in EBL (108 mL in the control vs. 54 mL in the KI group)

favoring the KI group. Whether this difference in EBL was

clinically significant is difficult to determine because this

study did not report other clinical outcomes, such as

complications. Moreover, the EBL among patients who

received KI in the study by Erbil et al. was similar to the

EBL among GD patients in our study who never received

KI. Thus, it may be the case that KI administration has an

effect that can also be achieved by operative technique in

the hands of experienced endocrine surgeons.

Our study has important limitations that must temper its

interpretation. As with any retrospective study, the data are

limited to that recorded by clinicians in the medical record.

The primary outcome, EBL, was determined from esti-

mates by surgeons recorded in the operative note rather

than precise measurements. Nevertheless, it is unlikely that

these estimates were systematically biased. Thus, the

imprecision inherent in these estimates would primarily

result in increased variance and a resultant decrease in

power. Even with this limitation, our study achieved

greater than 80 % power to detect a 40 mL difference in

blood loss between the two groups. It is unlikely that any

smaller difference would be clinically significant, even if a

higher-powered study could demonstrate a statistically

significant difference. Also, the ideal comparison group for

TABLE 1 Patient characteristics

Characteristics GD

(n = 162)

TMNG

(n = 102)

p

Mean age (years) 42.7 49.6 \0.001

Women 77 % 84 % 0.127

BMI classification 0.047

Underweight (\18.5 kg/m2) 5 % 1 %

Normal weight

(18.5–24.9 kg/m2)

28 % 23 %

Overweight

(25–29.9 kg/m2)

30 % 22 %

Obese (C30 kg/m2) 37 % 54 %

ASA classification 0.977

1 1 % 1 %

2 62 % 61 %

3 36 % 37 %

4 1 % 1 %

Mean mass of thyroid (g) 40.4 44.8 0.367

Mean length of follow-up

(months)

12.2 14.6 0.210

GD Graves disease, TMNG toxic multinodular goiter, BMI body mass

index, ASA American Society of Anesthesiologists

TABLE 2 Outcomes

Characteristics GD

(n = 162)

TMNG

(n = 102)

p

Mean EBL (mL) 55.4 51.5 0.773

Mean procedure time (min) 131.5 122.6 0.084

Transient hypocalcemia 51 (31 %) 50 (49 %) 0.004

Hypocalcemia [6 months 6 (4 %) 1 (1 %) 0.180

Transient RLN palsy 12 (7 %) 11 (11 %) 0.339

RLN paresis [6 months 1 (1 %) 2 (2 %) 0.432

Hematoma 1 (1 %) 2 (2 %) 0.325

Mean length of stay (days) 1.1 1.0 0.312

GD Graves disease, TMNG toxic multinodular goiter, EBL estimated

blood loss, RLN recurrent laryngeal nerve

2966 M. C. Shinall Jr. et al.

GD patients who did not receive KI would be GD patients

who did receive KI before TT by the same surgeons over

the same time period. Because so few patients receive KI at

our institution, such a comparison group was not available.

Nevertheless, TMNG patients represent a similar group of

patients for whom KI is not generally thought to benefit.

Demonstrating equivalent results between GD and TMNG

patients when not receiving KI preoperatively calls into

question its utility for GD patients.

Although KI is generally well tolerated, its administra-

tion nevertheless inconveniences patients and exposes

them to risk of adverse reactions. Beta-blockers and thi-

onamides have rendered obsolete KIs use in reducing

perioperative thyroid storm, its original justification. This

study adds to the literature calling into question the clinical

benefit of KIs effect in reducing thyroid gland vascularity,

the justification for its continued use. As thyroid surgery

continues to develop, it is perhaps time to rethink the

recommendation that KI routinely be administered to GD

patients before TT.

DISCLOSURES The authors declare no conflict of interest.

REFERENCES

1. Plummer HS. Results of administering iodine to patients having

exophthalmic goiter. JAMA. 1923;80:1955.

2. Bahn Chair RS, Burch HB, Cooper DS, et al. Hyperthyroidism

and other causes of thyrotoxicosis: management guidelines of the

American Thyroid Association and American Association of

Clinical Endocrinologists. Thyroid. 2011;21:593–646.

3. Werga-Kjellman P, Zedenius J, Tallstedt L, Traisk F, Lundell G,

Wallin G. Surgical treatment of hyperthyroidism: a ten-year

experience. Thyroid. 2001;11:187–92.

4. Barakate MS, Agarwal G, Reeve TS, Barraclough B, Robinson B,

Delbridge LW. Total thyroidectomy is now the preferred option

for the surgical management of Graves’ disease. ANZ J Surg.

2002;72:321–4.

5. Chiang FY, Wang LF, Huang YF, Lee KW, Kuo WR. Recurrent

laryngeal nerve palsy after thyroidectomy with routine identifi-

cation of the recurrent laryngeal nerve. Surgery. 2005;137:342–7.

6. Lal G, Ituarte P, Kebebew E, Siperstein A, Duh QY, Clark OH.

Should total thyroidectomy become the preferred procedure for

surgical management of Graves’ disease? Thyroid. 2005;15:569–

74.

7. Ku CF, Lo CY, Chan WF, Kung AW, Lam KS. Total thyroid-

ectomy replaces subtotal thyroidectomy as the preferred surgical

treatment for Graves’ disease. ANZ J Surg. 2005;75:528–31.

8. Weber KJ, Solorzano CC, Lee JK, Gaffud MJ, Prinz RA. Thy-

roidectomy remains an effective treatment option for Graves’

disease. Am J Surg. 2006;191:400–5.

9. Gaujoux S, Leenhardt L, Tresallet C, et al. Extensive thyroid-

ectomy in Graves’ disease. J Am Coll Surg. 2006;202:868–73.

10. Feliciano DV, Lyons JD. Thyroidectomy is optimal treatment for

Graves’ disease. J Am Coll Surg. 2011;212:714–20.

11. Hallgrimsson P, Nordenstrom E, Almquist M, Bergenfelz AO.

Risk factors for medically treated hypocalcemia after surgery for

Graves’ disease: a Swedish multicenter study of 1,157 patients.

World J Surg. 2012;36:1933–42.

12. Zambudio AR, Rodriguez J, Riquelme J, Soria T, Canteras M,

Parrilla P. Prospective study of postoperative complications after

total thyroidectomy for multinodular goiters by surgeons with

experience in endocrine surgery. Ann Surg. 2004;240:18–25.

13. Agarwal G, Aggarwal V. Is total thyroidectomy the surgical

procedure of choice for benign multinodular goiter? An evidence-

based review. World J Surg. 2008;32:1313–24.

14. Rienhoff WF. The histological changes brought about in cases of

exophthalmic goiter by the administration of iodine. Bull Johns

Hopkins Hosp. 1925;37:285–306.

15. Langley RW, Burch HB. Perioperative management of the thy-

rotoxic patient. Endocrinol Metab Clin N Am. 2003;32:519–34.

16. Ansaldo GL, Pretolesi F, Varaldo E, et al. Doppler evaluation of

intrathyroid arterial resistances during preoperative treatment

with Lugol’s iodide solution in patients with diffuse toxic goiter.

J Am Coll Surg. 2000;191:607–12.

17. Marigold JH, Morgan AK, Earle DJ, Young AE, Croft DN. Lu-

gol’s iodine: its effect on thyroid blood flow in patients with

thyrotoxicosis. Br J Surg. 1985;72:45–7.

18. Chang DC, Wheeler MH, Woodcock JP, et al. The effect of

preoperative Lugol’s iodine on thyroid blood flow in patients with

Graves’ hyperthyroidism. Surgery. 1987;102:1055–61.

19. Coyle PJ, Mitchell JE. Thyroidectomy: is Lugol’s iodine neces-

sary? Ann R Coll Surg Engl. 1982;64:334–5.

20. Marmon L, Au FC. The preoperative use of iodine solution in

thyrotoxic patients prepared with propranolol. Is it necessary? Am

Surg. 1989;55:629–31.

21. Yabuta T, Ito Y, Hirokawa M, et al. Preoperative administration

of excess iodide increases thyroid volume of patients with

Graves’ disease. Endocr J. 2009;56:371–5.

22. Erbil Y, Ozluk Y, Giris M, et al. Effect of lugol solution on

thyroid gland blood flow and microvessel density in the patients

with Graves’ disease. J Clin Endocrinol Metab. 2007;92:2182–9.

Potassium Iodide Solution in Graves Disease 2967