World J. Surg. 16, 654-662, 1992

World Journal of Surgery C 1992by the Soci6t~ lnternationale de Chirurgi©

Unilateral Neck Exploration for Primary Hyperparathyroidism: Analysis of a Controversy Using a Mathematical Model Quan-Yang Duh, M.D., Per U d r n , M.D., and Orlo H. Clark, M.D. Surgical Service, Veterans Affairs Medical Center and Department of Surgery, Mr. Zion Medical Center, University of California, San Francisco, California, U.S.A. Most endocrine surgeons explore both sides of the neck and identify all parathyroid glands when operating on patients with primary hyperparathyroidism. Others, however, advocate the unilateral approach, i.e., if an adenoma and a normal gland are identified, the contralateral side is not explored. We analyzed the strategy of the unilateral approach using a mathematical model to determine the variables that influence the probability of missing a tumor on the unexplored side of the neck. Assuming the frequency of single adenoma is 80%, hyperplasia 14%, double adenomas 4%, triple adenomas 1%, and carcinoma 1%, and the probability of missing a tumor on the explored side is 5%, we found that: 1. Only 41% of the patients treated by the unilateral approach undergo unilateral exploration. This is increased to 62% when a localization study with a sensitivity of 80% is used pre-operatively. 2. The probability of missing a tumor on the unexplored side of the neck parallels the prevalence of multiple adenomas. Half of the patients with triple adenomas and two-thirds of the patients with double adenomas will have a missed tumor when treated by the unilateral approach. 3. Patients who undergo unilateral exploration have an additional 7% to 8% probability of missing a tumor that would have been found if bilateral exploration is performed. This risk is lowered to 2% by a pre:operative localization study that is 80% sensitive. 4. A prospective study will require 684 patients, randomized to the unilateral or bilateral approach, to have an 80% statistical power (~ = 0.05,/3 = 0.20)of detecting a difference between a 5% and a 10% risk of missing a tumor.

that compare the outcome of these two approaches are retrospective, not randomized, and usually too small to have the statistical power to discern a small difference that may be clinically significant [5]. Contributing to this controversy is the question of whether multiple adenomas exist, or whether they are only extreme cases of asymmetrical hyperplasia [6]. A prospective randomized study has been proposed to resolve this controversy, but the number of patients required for a randomized trial appears prohibitive. In this paper we use a mathematical model to reduce this controversy of surgical approaches to variables that can be measured or assumptions that can be tested. We analyze how variables, such as the prevalence of multiple adenomas and the sensitivity of the pre-operative localization studies, influence the outcome, such as the frequency of actual unilateral exploration and the risk of missing an abnormal gland, when the unilateral approach is used. We also calculate the number of patients required for a prospective randomized study with enough statistical power to resolve this controversy. Method

The operative approach to patients with primary hyperparathyroidism is controversial. Most endocrine surgeons explore both sides of the neck and identify all parathyroid glands to avoid missing a parathyroid tumor that may cause persistent or recurrent hyperparathyroidism [1, 2]. Others, however, advocate unilateral exploration, that is, if an abnormal gland and a normal gland are identified on the initial side of exploration, the contralateral side is not explored [3, 4]. They argue that the morbidity of postoperative hypocalcemia and nerve injury is lower and that the incidence of persistent or recurrent disease is not higher when hyperparathyroidism is treated by the unilateral approach. This controversy has continued mainly because most studies Presented at the International Association of Endocrine Surgeons in Stockholm, Sweden, August, 1991~ Reprint request: Quan-Yang Duh, M.D., Veterans Affairs Medical Center, Surgical Service (112), 4150 Clement Street, San Francisco, California 94121, U.S.A.

A mathematical model for the unilateral approach was created using a spread sheet program (Microsoft Excel, Microsoft Corp., Redmond, Washington) on an IBM compatible micro" computer. The spread sheet model, in DOS or Mac format, is available from the corresponding author on request. The variables that were analyzed include the prevalence of multiple adenomas, the probability of missing a tumor on the side of the neck that is explored, and the sensitivity of a pre-operatiVe localization study for parathyroid tumors. The outcomes exa w° ined include the actual percentage of unilateral exploration when a unilateral approach is attempted, and the risk of missing an abnormal gland on the unexplored side of the neck. The outcome was calculated using values of variables that are generally accepted by the surgeons and are found in the surgical literature. How the outcome responds to the changes in the variables is also examined. The assumptions required for the analysis are explicitly stated in this model and can be examined, and the model can be altered by changing these assumptions'

Q.-Y. Dub et al.: Unilateral Parathyroidectomy

Definitions of the Surgical Approaches Unilateral exploration: One side of the neck is explored for the presence of normal and abnormal parathyroid glands. Bilateral exploration: Both sides of the neck are explored for the presence of normal and abnormal parathyroid glands. Unilateral approach: Unilateral approach refers to the intent to do a unilateral exploration if possible. One randomly chosen side of the neck is explored. If an abnormal parathyroid gland and a normal parathyroid gland are found, the contralateral side is not explored. If both glands are normal or both are abnormal, the contralateral side is explored. Unilateral approach only SOmetimes results in unilateral exploration; it frequently results in bilateral exploration. Bilateral approach: Bilateral approach refers to the intent to do a bilateral exploration. Both sides of the neck are explored for normal and abnormal parathyroid glands regardless of the finding on the initial side. Bilateral approach always results in bilateral exploration.

Variables The initial values were selected from the textbooks and literature. Analysis is then performed by changing these values.

1. Frequency of abnormal glands in patients undergoing initial neck exploration for primary hyperparathyroidism: single adenoma 80%, double adenomas 4%, triple adenomas 1%, hyperplasia 14%, carcinoma I%. 2. Probability of missing an abnormal gland on the side of the neck that is explored: 5%. 3. Sensitivity of a pre-operative localization study to correctly predict the side of the neck where the abnormal parathyroid gland can be found: 80%.

Assumptions 1. Four parathyroid glands are present in each patient. No supranumerary gland. 2. The abnormal glands are equally likely to be found on either side of the neck, and in the upper or the lower position. 3. A parathyroid gland will be missed if it is on the unexplored side of the neck. The gland can not be identified, for example, by palpation without a formal exploration. 4. An abnormal gland that is missed will cause persistent or recurrent disease. 5. The sensitivity of the pre-operative localization study is the same for each gland regardless of its location and regardless of whether the patient has a single adenoma or multiple adenomas. 6. The specificity of the pre-operative localization study is 100%, i.e., no false positive. Results

The outcome of the unilateral approach (percentage of actual unilateral exploration and the probability of missing an abnormal gland on the unexplored side) is first calculated assuming a 0% probability of missing an abnormal gland on the side of the

655

neck that is explored (Tables 1 and 2). If 100 patients are treated by the unilateral approach, 43 patients will actually undergo a unilateral exploration and 3.17 (7.34%) patients will have abnormal glands missed on the unexplored side of the neck. The other 57 patients will undergo bilateral exploration because either two normal glands or two abnormal glands are found on the initial side of exploration. The probability of finding one abnormal gland and one normal gland on the initial side of exploration in a patient with double adenomas is (1/2 x 2/3 + 1/2 x 2/3), or 2/3. The probability of finding one abnormal gland and one normal gland on the initial side of exploration in a patient with triple adenomas is (1/4 + 3/4 x 1/3), or 1/2. Half to two-thirds of the patients with multiple adenomas will undergo unilateral exploration and, therefore, will have a tumor missed on the unexplored side of the neck. The outcome of the unilateral approach is then calculated assuming a 5% probability of missing an abnormal gland on the side that is explored (Table 3). If 100 patients are treated by the unilateral approach, 41 patients will actually undergo a unilateral exploration, and 3. I (7.59%) patients will have abnormal glands missed on the unexplored side of the neck. The other 59 patients will undergo bilateral exploration. When the probability of missing an abnormal gland on the explored side (variable P) is increased from 0% to 25%, the percentage of patients undergoing unilateral exploration decreased, as expected, from 43% to 33%. The percentage of patients who have a missed tumor in the unexplored side, however, increases only minimally from 7.34% to 8.57% (Table 4). The probability of missing a tumor on the unexplored side is not influenced significantly by the probability of missing a tumor on the explored side. When the percentage of multiple adenomas is increased from 0% to 25%, the perecentage of patients undergoing unilateral exploration increased from 38% to 52% (Table 5). The probability of missing a tumor on the unexplored side increases in parallel with the prevalence of multiple adenomas from 0% to 31% (Table 5). The probability of missing a tumor on the unexplored side in a patient undergoing unilateral exploration is approximately equal to the probability of the patient having multiple adenomas. The outcome is again calculated, assuming that the patient has undergone a pre-operative localization study. When the sensitivity of the pre-operative study is increased from 50% to 95%, the percentage of patients undergoing unilateral exploration is increased from 41% to 73% and the percentage of patients with missed tumors in the unexplored side is decreased from 6.8% to 0.5% (Tables 6 and 7). The more sensitive the pre-operative localization study, the more likely the patient will undergo unilateral exploration, and the less likely a tumor will be missed in the unexplored side of the neck. Discussion

The intent to perform a unilateral exploration, i.e., the unilateral approach, is not the same as unilateral exploration. Without a pre-operative localization study, only about 40% of the patients who are treated by the unilateral approach actually undergo unilateral exploration (Table 2). Even with a preoperative localization study that is 80% sensitive and 100%

656

World J. Surg. Vol. 16, No. 4, July/Aug. 1992

Table 1. Formulas for calculating the percentages o f patients undergoing unilateral exploration and percentage o f these patients with missed tumors.a % patients

Pathology

With the pathology

Explore one side

Explore both sides

With tumors missed on the unexplored With tumors missed side on the explored side

With tumor missed if both sides were explored routinely

One a d e n o m a

AI

C, = At - B,

DI = 0

Et = Ai × P

F1 = At × P

Two adenomas Initial side o f exploration One

A2

B, = 1/2 × A t x (l-P) B2 = B2i + B2ii

C2 = C2i + C2ii

D2 = D2i + D2ii

E2 = E2i + E2ii

F2 = A2 x ( 1 - ( l - P ) 2)

B2i = A2i x (i-P) B~ii = A:ii × P × (l-P)

C2i = A2i - B2i

Dzi = Bzi

E2i=A2i x P=C2i

Both or none

Three adenomas Initial side o f exploration One Two Hyperplasia Carcinoma Total

A2i = 2/3 × A~ A:ii = 1/3 × A:

C2ii = A2ii - B2ii D2ii = B2ii

E2ii = Azii × P = 1/2 × A2ii x PZ + I/2 x Azii x ( I - ( I p)2)

Ca = C3i + C3ii

D3 = D3i + D3ii

E~ = E3i + E3ii

D3i = B3i

E~i= A~i× P=C3i

D3ii = B3ii

E3ii = A3ii × (p2 + (I-P) z × P)) E4 = A4 × (1-(I-P) 4) F4 = A4 × (I--(I-P) 4) E 5 = A5 x P F ~ = A5 × P E = 5'. E(I_5) F = Sum F(I_5)

A3

B3 - B3i + B3ii

A3i = 1/2 × A3

B3i = A3i x C3i = Aai - B3i (l-P) B3ii = A3ii × 2P Czii = A3ii - B3ii x (l-P) B4 = 0 (24 = A4 Bs = 0 C5 = As B = X B(i-5) C = X C._~)

A3ii = 1/2 x A3 A4 As A = .X A(l_~

D4 = 0 D~ = 0 D = X D(I_s)

F3 = A3 × (1-(I-P) 3)

OThe probability o f missing a tumor even if the side o f the neck is explored is P. The probability o f having only one a d e n o m a on the initial side o f exploration in a patient with double adenomas is A2i = (1/2 x 2/3 + 1/2 x 213) × A 2 = 2/3 x A2, and in a patient with triple a d e n o m a s is A 3 i = ( 1 / 4 + 3/4 x 1/3) x A 3 = 1/2 × A 3. Calculated difference b e t w e e n the unilateral approach and bilateral approach: Single adenoma, D l + E 1 - F t = 0; Double adenomas, D: + E 2 - F 2 = A 2 × (2/3 - 1 1/3 P + 2/3 p2); Triple adenomas, D 3 + E 3 - F 3 = A 3 x (1/2 - 1 1/2 P + 1 1/2 pz _ 1/2 p3); Hyperplasia, D 4 + E4 F4 = 0; and Carcinoma, D5 + Es - F~ = 0.

Table 2. Comparison o f the frequency o f missed tumors by the unilateral approach versus the bilateral approach." % patients

Pathology One a d e n o m a Two adenomas Initial side o f exploration One Both or none Three a d e n o m a s Initial side o f exploration One Two Hyperplasia Carcinoma Total

With the pathology 80 4 2.67 1.33 1 0.5 0.5 14 1 100

Explore one side

Explore both sides

With tumors missed on the unexplored side

With tumors missed on the explored side

With tumor missed if both sides were explored routinely

40 2.67

40 1.33

0 2.67

0 0

0 0

2.67 0 0.5

0 1.33 0.5

2.67 0 0.5

0 0 0

0

0.5 0 0 0 43.17

0 0.5 14 ! 56.83

0.5 0 0 0 3. t7

0 0 0 0 0

0 0 0

~The probability o f missing a tumor even if the side o f the neck is explored, P, is assumed to be 0. The probability o f double adenomas is 4% and triple a d e n o m a s is 1%.

s p e c i f i c , o n l y 6 2 % o f p a t i e n t s will a c t u a l l y u n d e r g o u n i l a t e r a l exploration. When designing a comparative study, one should not equate unilateral approach with unilateral exploration. Patients can be randomly assigned to the unilateral approach or the bilateral approach when undergoing parathyroidectomy, but they cannot b e r a n d o m l y a s s i g n e d to u n i l a t e r a l e x p l o r a t i o n o r b i l a t e r a l e x p l o r a t i o n . W h e t h e r a p a t i e n t is t r e a t e d w i t h t h e u n i l a t e r a l

approach can be determined pre-operatively; whether a patient a c t u a l l y u n d e r g o e s u n i l a t e r a l e x p l o r a t i o n d e p e n d s o n t h e intraoperative findings. It is e r r o n e o u s t o c o m p a r e p a t i e n t s w h o h a v e h a d u n i l a t e r a l e x p l o r a t i o n w i t h t h o s e w h o h a v e h a d b i l a t e r a l e x p l o r a t i o n to d e t e r m i n e w h e t h e r t h e u n i l a t e r a l a p p r o a c h is s u p e r i o r o r inferior to t h e b i l a t e r a l a p p r o a c h . A m o n g t h e p a t i e n t s t r e a t e d b y t h e unilateral approach, the ones who undergo bilateral exploration

657

Q.'Y, Duh et ai.: Unilateral Parathyroidectomy Table 3. Comparison of the frequency of missed tumors by the unilateral approach versus the bilateral approach, a % patients

Pathology One adenoma Two adenomas Initial side of exploration One Both or none Three adenomas Initial side of exploration One Two cHyperplasia arcinoma Total

With the pathology 80 4 2.67 1.33 1 0.5 0.5 14 1 100

Explore one side

Explore both sides

With tumors missed on the unexplored side

With tumors missed on the explored side

With tumor missed if both sides were explored routinely

38 2.60

42 1.40

0 2.60

4 O. 14

4 0.39

2.53 0.063 0.52

0.13 1.27 0.48

2.53 0.063 0.52

0.013 0.12 0.068

0.14

0.475 0.0475 0 0 41,12

0.025 0.45 14 I 58.88

0.475 0.0475 0 0 3.12

0.0036 0.065 2.60 0.05 6.85

2.60 0.05 7.18

#The probability of missing a tumor even if the side of the neck is explored, P, is assumed to be 5%. The probability of double adenomas is 4% and triple adenomas is 1%. Table 4. Percentage of unilateral exploration and percentage of missed tumors when the unilateral approach is attempted, calculated using Table 1.~ % of patients Probability of missing Who actually a tumor on the side undergo Undergoing unilateral of the neck that is unilateral exploration who have explored exploration a missed tumor 0 43. t7 7.34 0.05 41.12 7.59 0.10 39.06 7.83 0,15 36.99 8.08 0.20 34.91 8.33 0.25 32.81 8.57 ""OThe probability of missing an abnormal gland on the explored side, P, is varied from 0% to 25%. The probability of double adenomas is 4% and triple adenomas is 1%..

Can have a higher probability of a missed tumor than those who Undergo unilateral exploration (Table 3), Since one always explores the contralateral side when the tumor is missed on the initial side, patients who have a tumor that is difficult tO find are taore likely to undergo bilateral exploration. This accounts for the apparent paradox of a higher rate of missed tumor in Patients who have undergone bilateral exploration than those Who have undergone unilateral exploration [7]. Without a pre-operative localization study, the probability of nlissing a tumor in the unexplored side o f the neck is about 8% (Table 1). This probability is not significantly influenced by the Probability o f missing a tumor on the explored side (Table 4). The 8% tumors missed are not tumors that are difficult to find, but tumors that would have been found if both sides of the neck Were explored. The prevalence of multiple adenomas determines the probability of missing a tumor on the unexplored side in the patient Undergoing unilateral exploration. H a l f to two-thirds o f the Patients with multiple adenomas will have a missed tumor on the unexplored side o f the neck (Table 1). This probability of missing a tumor approximates the prevalence o f multiple adenOmas (Table 5). Thus, a population that has a higher preva-

Table 5. Percentage of actual unilateral exploration and percentage of missed tumors when the unilateral approach is attempted, calculated using Table 1.4 % of patients

With 2 adenomas 0 1 2 2.2 3 4 5 10 15 20 25

With hyperptasia 19 18 17 16.8 16 15 14 9 4 0 0

Who actually undergo unilateral exploration 38 38.65 39.3 39.43 39.95 40.6 41.25 44.49 47.74 50.98 51.85

Undergoing unilateral exploration who have a missed tumor 0 1.68 3.30 3.62 4.88 6.40 7.87 14.59 20.40 25.47 31.30

aThe probability of missing an abnormal gland on the explored side, P, is assumed to be 5%. The percent of double adenomas is varied from 0% to 25%. The percent of triple adenoma is 0%. The percent of carcinoma is 1%.

lence of multiple adenomas is more likely to have a missed tumor in the unexplored side. The literature shows a wide range o f prevalence o f multiple adenomas, from 0% to 10%, averaging 2.2% (Table 8). This difference in the prevalence of multiple adenomas is most likely caused by the different techniques used to determine multiple adenomas versus hyperplasia by different surgeons and pathologists, Some surgeons excise or biopsy the normal appearing gland for a frozen section and Oil-Red-O staining on the initial side when a tumor is found so as not to miss a deceptively small hyperplastic gland. They, therefore, diagnose hyperplasia more often and multiple a d e n o m a s less often. Other surgeons may rely only on the size o f the gland to determine abnormality. They, therefore, diagnose more multiple adenomas and fewer hyperplasia. The former will have a lower risk o f missing a

658

World J. Surg. Voi. 16, No. 4, July/Aug. 1992

Table 6. Formulas for calculating the percentages of unilateral exploration and missed tumors. ° % patients

Pathology

With the pathology

Unilateral exploration

Bilateral exploration

With tumors missed on the With tumors missed unexplored side on the explored side

With tumors missed if both sides routinely explored

Oneadenoma Two adenomas

At A2

BI = A t x Q B2 = Z B2(i-iii)

C 1 = Aj = B 1 C 2 = Z Cz(i-iii)

Di = 0 Et = C 1 x P D2 = .X D2(i-iii) E2 = X E2(i-iii)

F) = A 1 x (l-Q) x P Fz = A2 x(l--(l-(1-Q) x p)2)

A2i = A2 x (l--Q) 2 Azii = A2 × 2Q × (I-Q) A2iii = A2 x Q2

B2i = A2i x ((2t3 Czi = A2i = Bzi Dzi = Bzi x (l-P) + 1/3 x P x (I-P)) B2ii = A2ii x (2/3 Cuii= Azii = B2ii D2ii = Bzii + 1/3 P) D2iii = 0 B2iii = 0 C2iii = A2iii

A3

B 3 = X B3(Mv)

C3 = X C3(i-iv)

D3 = E D3(i-iv) E 3 = Z E3(i-iv)

B3i = A3i (1/2 x (l-P) + P x (l-P)) A3ii = A3 x 3Q Bfii = A3ii (1/3 x (I-Q) 2 + 2/3 x P) A3iii -- A 3 x 3Q2 B3iii = 0 x (l--Q) A3iv = A 3 x Q3 B3iv = 0

C3i = A3i = B3i

D3iv = B3iv

Study locates none Study locates one Study locates both Three adenomas Study locates none Study locates one Study locates two Study locates three

A3i = A 3 x (l-Q) 3

C3ii = A3ii = B3ii D3ii = B3ii

Exi = Azi x P Ezii = 0 E2iii = 0 F3 = A3 x (1-O-0-Q) x p)3)

E3i = A3i x (1/2 x (P + p2 + (l_p)2 x P) E3ii = Cfii x P

C3iii = A3iii

D3ii = 0

E3iii = C3iii x P

C3iv = A3iv

Dfiii = 0

Eziv = 0 E 4 = C 4 x (1--(1-(I-Q) F 4 = A 4 X (1--(1-(I-Q) x p)4) x p)4) E s = A 5 x (l-Q) x P F 5 = A s x (I-Q) x P E = Z E._5) F = X F(1-5)

Hyperplasia

m4

B4 = 0

C4 = A4

D4 = 0

Carcinoma Total

A5 A = E A~,_s)

B~ = 0 B = X Bu_~)

C 5 = A~ C = ~. C._5)

D5 = 0 D = X D(I_~

aThe probability of missing a tumor even if the side of the neck is explored is P. The probability of a pre-operative localization study correctly predict the side of the tumor is Q. P and Q are independent of each other.

Table 7. Calculated percentage of patients undergoing unilateral exploration and the percent of these patients with missed tumors when the unilateral approach is attempted, using Table 6. a % probability that the pre-operative study correctly locate the tumor 50 60 70 80 85 90 95

% of patients undergoing

Unilateral exploration

Exploration who have a missed tumor on unexplored side

40.76 47.98 55.12 62.16 65.65 69.12 72.57

6.77 4.97 3.48 2.19 1.60 1.04 0.51

aThe probability of missing an abnormal gland on the explored side, P, is assumed to be 5%. The percent of double adenomas is 4%, triple adenoma is 1%, hyperplasia is 14%, and carcinoma is 1%. The probability that the preoperative localization study correctly localize a tumor is varied from 50% to 95%.

t u m o r in the u n e x p l o r e d side o f the neck than the latter, w h e n the unilateral a p p r o a c h is used. Our analysis does not d e p e n d on h o w multiple a d e n o m a s are diagnosed. The p r e v a l e n c e of multiple a d e n o m a s d e t e r m i n e d by a surgeon in the patient population should be u s e d to d e t e r m i n e the risk of missing a t u m o r if a unilateral a p p r o a c h is used (Table 5). T h e unilateral a p p r o a c h will be m o r e likely to miss a t u m o r in a population that has a higher p r e v a l e n c e o f multiple a d e n o m a s , for e x a m p l e , patients with multiple e n d o c r i n e neoplasia ( M E N ) and familial h y p e r p a r a t h y r o i d i s m . In our recent r e v i e w o f 250

patients with primary h y p e r p a r a t h y r o i d i s m treated by the bilateral a p p r o a c h , the p r e v a l e n c e of multiple a d e n o m a s is significantly higher in patients older than 60 years (10%) than in patients y o u n g e r than 60 years. This higher p r e v a l e n c e o f multiple a d e n o m a s in older patients has b e e n r e p o r t e d previously [40]. T h e older patients m a y be at a higher risk for a failed p a r a t h y r o i d e c t o m y if the unilateral a p p r o a c h is used. A sensitive p r e - o p e r a t i v e localization study increases the probability of patients actually u n d e r g o i n g unilateral exploration and d e c r e a s e s the probability o f missing a t u m o r in the u n e x p l o r e d side. T h e m o r e a c c u r a t e the p r e - o p e r a t i v e study, the more likely the patient will u n d e r g o unilateral e x p l o r a t i o n and the less likely a t u m o r will be missed in the u n e x p l o r e d side o f the n e c k (Table 7). A sensitive localization study i m p r o v e s the o u t c o m e by identifying s o m e patients with multiple adenomas and excluding t h e m from a unilateral exploration. T h e available parathyroid localization studies, ultrasonography, c o m p u t e r i z e d t o m o g r a p h y , m a g n e t i c r e s o n a n c e imaging, and thallium-technetium scanning, all h a v e a sensitivity of about 80%. T h e y are usually not r e c o m m e n d e d b e f o r e the initial neck exploration b e c a u s e of the cost and minimal utility w h e n bilateral exploration is d o n e [41]. W h e t h e r it is cost effective to use localization studies routinely in patients u n d e r g o i n g initial n e c k e x p l o r a t i o n to i m p r o v e the o u t c o m e o f the unilateral a p p r o a c h is debatable [5, 42]. O n e can calculate the n u m b e r o f patients r e q u i r e d for a p r o s p e c t i v e r a n d o m i z e d study to c o m p a r e the o u t c o m e o f the unilateral a p p r o a c h to that o f the bilateral a p p r o a c h . B e c a u s e o n l y 40% o f the patients r a n d o m i z e d to the unilateral a p p r o a c h

Q.-Y. Duh et al.: Unilateral Parathyroidectomy

659

Table 8. Incidence of multiple parathyroid adenomas reported in the literature.

-......

Authors

---.._.

Woolner et al. Black Hellstr6m and Ivemark Cope Block et al. Haft et al.

Year 1952 1961 1962

No. of patients with multiple % of patients adenomas/tota! with multiple Reference no. of patients adenomas 8 3/137 2.1 9 11/385 2.8 10 3/138 2. I

1966 1967 1970 1973 1973 1974 1974 1976 1976 1976

1! 12 13 14 15 16 17 18 19 20

Paloyan et al. Romanus et al. Esselstyn et al. Myers Castleman et al. Clark et al. Paloyan and Lawrence 1977 21 Coffey et aL 1977 22 Cooke et al. Wang et aL 1977 23 1978 24 Block et aL Harness et al. 1979 6 1981 25 Carnevale et al. Verdonk and Edis 1981 26 1982 27 I-lines et al. Johansson et al. 1982 28 Russel and Edis 1982 29 1982 30 Thompson et aL Paloyan et al. 1983 31 1983 32 Trigonis et al. Farnebo et al. 1984 33 1985 34 Bondes0n et al. Wang 1085 4 Brothers and 1987 35 Thompson 1987 36 Gaz et al. Roses et al. 1989 37 Wallfelt et aL 1990 38 1990 39 Attic et aL Total

10/200 3/70 0/47 0/84 6/274 0/111 3/82 7/557 13/295 9/180

5.0 4.3 0 0 2.0 0 3.7 1.3 4.4 5.0

4/200 2/102 0/73 0/206 5/300 4/38/1962 7/176 0/208 14/472 7/273 0/292 8/263 15/400 0/191 6/1000 7/81

2.0 2.0 0 0 1.7 1.9 4.0 0 3.0 2.6 0 3.0 3.7 0 0.6 9.0

0/50 3/105 5/570 33/865 266/10429

0 2.9 0.9 3.2 2.2

Will undergo a unilateral exploration, the number o f patients required for such a study is large. For example, if bilateral approach is expected to miss 5% of the tumors and unilateral approach 10%, 342 patients will be required for each arm of the Study to have a 80% chance (statistical power) to show this difference, with a a of 0.05 and /3 of 0.20 [43]. ~ is the Probability of a type I error, i.e., the probability that we find a difference between the two groups when no difference exists./3 is the probability of a type II error, i.e., the probability that we find no difference between the groups when a difference actually exists. A smaller sample size would risk a higher type II error. The number of patients required for this study increases if the chance of missing a tumor is higher than 5% in the explored side (because fewer patients will have unilateral exploration), or if the expected difference between the two groups is smaller than 5% (e.g., 2.8% from Table 3). Whether unilateral approach is a good strategy depends on the additional benefit versus the additional risk of exploring the Contralateral side when an adenoma is already found. The risk of not exploring the contralateral side is the possibility of

missing a parathyroid tumor that can cause persistent or recurrent hyperparathyroidism. This risk is 7.9% when the prevalence of double adenomas is 5%, and 3.6% when the prevalence of double adenomas is 2 . 2 % . The risks o f exploring the contralateral side include nerve injury and hypoparathyroidism. Whether it is more reasonable to explore the contralateral neck for a 3.6% to 7.9% possibility of a missed gland, or to re-operate on these patients when the initial operation fails, depends on the surgeon and the expected complication rate. The bilateral approach is favored if the rate of complication is low, whereas the unilateral approach is favored if the rate of complication is high. I n summary, in this analysis we emphasized the distinction between unilateral approach and unilateral exploration. We showed that the probability of missing a parathyroid tumor on the unexplored side of the neck is primarily determined by the prevalence of multiple adenomas, and that half to two-thirds of the patients with multiple adenomas will have a missed tumor on the Unexplored side of the neck. Patients who are more likely to have multiple adenomas, such as the elderly, may be at higher risk for missed tumors. A sensitive pre-operative localization study is useful for the unilateral approach, because it increases the probability of unilateral exploration and decreases the r i s k o f a missed tumor. R6sum6 La plupart des chirurgiens explorent la rrgion cervicale de fa~on bilatrrale et identifient toute les glandes parathyroides au cours de l'intervention pour hyperparathyroidie primaire. D'autres, toutefois, plaident pour un abord unilat6ral et, lorsqu'un adrnome et une glande normale sont identifirs, n'explorent pas la rrgion controlatrrale. Nous avons analys6 la stratrgie de l'abord unilatrral ~ l'aide d'un modrle mathrmatique afin de drterminer les variables qui augmentent la probabilit6 de mrconnaRre une tumeur dans la rrgion cervicale inexplorre. En estimant les frrquences d'ad6nome unique 80%, d'hyperplasie ~t 14%, de double adrnome ~t 4%, de triple adrnome ~ 1% et de carcinome ~ 1% ainsi que la probabilit6 de m6conna/tre une tumeur dans la rrgion inexplorre/l 5%, nous avons trouv6 que: 1) seuiement 41% des patients traitrs par abord unilatrral subissent une exploration unilatrrale. Ce pourcentage passe/~ 62% quand une exploration topographique avec une sensibilit6 de 80% est pratiqure en prrop6ratoire. 2) la probabilit6 de mrconnaitre une tumeur de la r6gion cervicale non explorre est parallrle/~ la prrvalence des adrnomes multiples. La moiti6 des patients ayant un triple ad6nome et 2/3 des patients ayant un double adrnome peuvent avoir une tumeur m6connue s'ils sont traitrs par un abord unilat6ral. 3) les patients qui brn~ficient d'une exploration unilat6rale ont une probabilit6 augmentre de 7 ~ 8% d'avoir une tumeur mrconnue qui aurait 6t6 trouvre si une exploration bilatrrale avait 6t6 rratisre. Ce risque diminue h 2% lorsqu'une exploration topographique d'une sensibilit6 de 80% est rralisre en pr~oprratoire. 4) une 6tude prospective randomisre unilat6rale ou bilat6ral nrcessiterait 684 patients pour avoir une puissance statistique de 80% (a = 0,05,/3 = 0,20) pour mettre en 6vidence une diffrrence entre les risques de mrconnaitre une tumeur de 5% et de 10%.

660 Resumen La mayorfa de los cirujanos endocrinos expioran ambos lados del cuello e identifican todas las paratiroides cuando intervienen pacientes con hiperparatiroidismo primario. Otros, sin embargo, preconizan el abordaje unilateral: cuando se identifican un a d e n o m a y una gMndula normal no se explora el otro lado. Por ello nos propusimos analizar el abordaje unilateral utili. zando un modelo matem~ftico para determinar las variables que influencian la probabilidad de no detectar un tumor en et lado no explorado. Asumiendo que la frecuencia de un a d e n o m a tinico es 80%, de hiperplasia 14%, de adenomas dobles 4%, de adenomas triples 1%, de carcinoma 1%, y 5% la probabilidad de no detectar un tumor en el lado que se explora, encontramos que: 1. En s61o 41% de los pacientes tratados mediante el abordaje unilateral se efecttla exploraci6n unilateral, frecuencia que aument6 a 62% cuando se utiliz6 un estudio preoperatorio de localizaci6n con sensibilidad de 80%. 2. L a probabilidad de no detectar un tumor en el lado no explorado es igual a ia de los adenomas mtlltiples. L a mitad de los pacientes con adenoma triples y dos tercios de los pacientes con adenomas dobles resultan con un tumor no detectado cuando se realiza el abordaje unilateral. 3. Los pacientes sometidos a exploraci6n unilateral tienen una probabilidad adicional de 7 a 8% de no detecci6n de un tumor que habrfa sido identificado con una exploraci6n bilateral. Tal riesgo se disminuye a 2% mediante un estudio preoperatorio de localizaci6n que posea 80% de sensibilidad. 4. Un estudio prospectivo requiere disponer de 684 pacientes para ser randomizados al abordaje unilateral o al bilateral para lograr un potencial estadfstico de 80% (a = 0.05, /3 = 0,20) en la detecci6n de una diferencia de 5% a 10% en el riesgo de no identificar la presencia de un tumor. Acknowledgments We would like to thank Dr. Sten Tibblin for reviewing the manuscript. Supported in part by the Medical Research Service of the Veterans Affairs Medical Center, San Francisco, California, U.S.A. References 1. Thompson, N.W.: Localization studies in patients with primary hyperparathyroidism. Br. J. Surg. 75:97, 1988 2. Clark, O.H., Duh, Q.Y.: Primary hyperparathyroidism: A surgical perspective. Endocrinol. Metab. Clin. North Am. 18:701, 1989 3. Tibblin, S., Bondeson, A.G., Bondeson, L., Ljungberg, O.: Surgical strategy in hyperparathyroidism due to solitary adenoma. Ann. Surg. 200:776, 1984 4. Wang, C.: Surgical management of primary hyperparathyroidism. Curr. Probl. Surg. 12:1-502, 1985 5. Lucas, R.J., Welsh, R.J., GIover, J.L.: Unilateral neck exploration for primary hyperparathyroidism. Arch. Surg. 125:982, 1990 6. Harness, J.K., Ramsburg, S.R., Nishiyama, R.H., Thompson, N.W.: Multiple adenomas and the parathyroids: Do they exist? Arch. Surg. 114:468, t979 7. Wang, C.A.: Invited commentary: Unilateral neck exploration for primary hyperparathyroidism. Arch. Surg. 125:985, 1990 8. Woolner, L.B., Keating, F.R., Black, B.M.: Tumors and hyperplasia of the parathyroid glands. Cancer 5:i069, 1952 9. Black, B.M.: The pathology and surgery of the parathyroid glands. In The Parathyroids: Proceedings of a Symposium on Advances in Parathyroid Research, R.O. Greep, R.V. Talmage, editors, Springfield, Illinois, Thomas Publ., 1961, p. 427

World J. Surg. Vol. 16, No. 4, July/Aug. 1992 10. Hellstrrm, J., Ivemark, B.I.: Primary hyperparathyroidism: Clinical and structural findings in 138 cases. Acta Chir. Scand. (Suppl.) 294:66, 1962 11. Cope, O.: The story of hyperparathyroidism at the Massachusetts General Hospital. N. Engl. J. Med. 274:1174, 1966 12. Block, M.A., Greenawald, K., Horn, R.C., Frame, B.: Involvement of multiple parathyroids in hyperparathyroidism. Am. J. Surg. 114:530, 1967 13. Haft, R.C., Black, W.C., Ballinger, W.F.: Primary hyperparathyroidism: Changing clinical, surgical and pathologic aspects. Ann. Surg. 171:85, 1970 14. Paloyan, E., Paloyan, D., Pickleman, J.R.: Hyperparathyroidisrn today. Surg. Clin. North Am. 53:211, 1973 15. Romanus, R., Heimann, P., Niisson, O., Hansson, G.: Surgical treatment of hyperparathyroidism. Prog. Surg. •2:22, 1973 16. Esselstyn, C.B. Jr., Levin, H.S., Eversman, J.E., Schumacher, O.P., Skillern, P.G.: Reappraisal of parathyroid pathology in hyperparathyroidism. Surg. Clin. North Am. 54:443, 1974 17. Myers, R.T.: Follow up study of surgically-treated primary hyperparathyroidism. Ann. Surg. 179:729, 1974 18. Castleman, B., Schantz, A., Roth, S.I.: Parathyroid hyperplasia in primary hyperparathyroidism. Cancer 38:1668, 1976 19. Clark, O.H., Way, L.W., Hunt, T.K.: Recurrent hyperparathyroidism. Ann. Surg. 184:391, 1976 20. Paloyan, E., Lawrence, A.M.: The rationale for subtotal parathyroidectomy. In Controversy in Surgery, R.L. Varco, J.P. Delaney, editors, Philadelphia, W.B. Saunders, 1976, pp. 18--72 21. Coffey, R.J., Lee, T.C., Canary, J.J.: The surgical treatment of primary hyperparathyroidism: A twenty year experience. Ann. Surg. 185:518, 1977 22. Cooke, T.J.C., Boey, J.H., Sweeney, E.C., Gilbert, J.M., Taylor, S.: Parathyroidectomy: Extent of resection and late results. Br. JSurg. 64:153, 1977 23. Wang, C.A.: Parathyroid reexptoration: A clinical and pathological study of 112 cases. Ann. Surg. 186:140, 1977 24. Block, M.A., Frame, B., Jackson, C.E.: The efficacy of subtotal parathyroidectomy for primary hyperparathyroidism due to multiple gland involvement. Surg, Gynecol. Obstet. 147:!, 1978 25. Carnevate, N., Samson, R., Bennett, B.P.: Multiple parathyroid adenomas. J.A.M.A. 246:1332, 1981 26. Verdonk, C.A., Edis, A.J.: Parathyroid "double adenomas": Fact or fiction? Surgery 90:523, 1981 27. Hines, J.R., Atiyah, R., Kliefoth, J., Beal, J.M.: Hyperparathyroidism: Problems in surgical management. Am. J. Surg. 144:504, 1982 28. Johansson, H., Thorrn, L., Werner, I.: Hyperparathyroidism: Clinical experiences from 208 cases. Upsala Med. Sci. 77:41, t982 29. Russell, C.F.;~Edis, A.J.: Surgery for primary hyperparathyroidism: Experience of 500 consecutive cases and evaluation of the role of surgery in the asymptomatic patient. Br. J, Surg. 69:244, 1982 30. Thompson, N.W., Eckhauser, F.E., Harness, J.K.: The anatomy of primary hyperparathyroidism. Surgery 92:814, 1982 31. Paloyan, E., Lawrence, A.M., Oslapas, R., Shah, K.H., Ernst, K., Hofmann, C.: Subtotal parathyroidectomy for primary hyperparathyroidism. Arch. Surg. 118:425, 1983 32. Trigonis, C., Hamberger, B., Farnebo, L.O., Abarca, J., Granberg, P.O.: Primary hyperparathyroidism: Changing trends over 50 years. Acta Chir. Scand, 149:675, 1983 33. Farnebo, L.O., Trigonis, C., Forsgren, L., Granberg, P.-O., Hamberger, B.: Surgery for primary hyperparathyroidism. Acta Chir. Scand. (Suppl.) 520:11, 1984 34. Bondeson, A,G., Bondeson, L., Ljungberg, O,, Tibblin, S.: Surgical strategy in nonfamilial primary parathyroid hyperplasia: Longtei'm follow-up of thirty-nine cases. Surgery 97:569, 1985 35. Brothers, T,E., Thompson, N.W.: Surgical treatment of primary hyperparathyroidism in elderly patients. Acta Chir. Scand. 153:175, I987 36. Gaz, R.D., Doubler, P.B., Wang, C.: The management of 50 unusual hyperfunctioning parathyroid glands. Surgery 102:949, 1987 37. Roses, D.F., Karp, N.S., Sudarsky, L.A., Valensi, Q.J., Rosen, R.J., Blum, M.: Primary hyperparathyroidism associated with two enlarged parathyroid glands. Arch. Surg. t24:1261, 1989 38. Wallfelt, C., Ljunghall, S., Bergstrrm, R., Rastad, J., Akerstrrm,

Q.-Y. Duh et al.: Unilateral Parathyroidectomy

661

G.: Clinical characteristics and surgical treatment of sporadic primary hyperparathyroidism with emphasis on chief cell hyperplasia. Surgery 107:13, 1990 39. Attie, J.N., Bock, G.B., Auguste, L.J.: Multiple parathyroid adenomas: Report of thirty-three cases. Surgery 108:1014, 1990 40. Thomson, N.W.: Discussion: Unilateral neck exploration for primary hyperparathyroidism. Arch. Surg. 125:985, 1990

41. Duh, Q.Y., Sancho, J.J., Clark, O.H.: Parathyroid localization. Acta Chir. Scand. 153:241, 1987 42. Uden, P., Aspelin, P., Berglund, J., Lilja, B., Nyman, U., Olsson, L.E., Zederfeldt, B.: Preoperative localization in unilateral parathyroid surgery. Acta Chir. Scand. 156:29, 1990 43. Zar, J.H.: Biostatistical analysis. Englewood Cliff, New Jersey, Prentice-Hall, 1984, pp. 397-400

Invited Commentary

often inappropriately includes patients who have had an unrewarding ipsilateral search who go on to have a contralateral exploration. This subgroup of bilateral exploration patients contains more cases with hard to find diseased glands in unusual places that may not be identified with resultant unsuccessful outcomes. The assumptions used in the paper are well founded, using pathologic diagnostic incidence ranges and surgical success rates which are generally accepted for centers with endocrine surgical expertise. Applying these statistics to the "average" surgeon performing parathyroid surgery must be adjusted to take into account lower success rates and higher complication rates. The biological variability of human beings and the disparate methods of surgeons makes one wary of the general conclusions of a paper based on theoretical calculations without hard data from the real world of clinical correlation with actual patients, operations, complications, success and failure rates, and detailed cost benefit outcome analysis. But such a prospective, randomized, controlled trial with adequate numbers of patients will probably never be done. Nevertheless, the calculations of failure rates (7%) for the unilateral approach are supported by the similar data (4%) of expert surgeons who espouse the unilateral approach [1]. It should not be assumed that a bilateral exploration guarantees finding all diseased and normal glands. It does seem logical that a more extensive search, whether it is unilateral or bilateral is more apt to discover missing or supernumerary hyperfunctioning glands. The political pendulum of parathyroid procedures swings from the "liberal" approach of subtotal 3 and V2gland parathyroidectomy [2] to the "conservative" unilateral approach [3]. Perhaps the "moderate" middle ground bilateral approach is that which is most popular and will continue to be accepted by most surgeons. Which of these approaches seems most radical is determined by the surgicopolitical perspective of the analyst. Is there a role for the selective use of unilateral explorations? With high operative or anesthetic risk patients, in the unstable hypercalcemic crisis patient in extremis, in cases performed under local anesthesia, or neck reoperations such as following prior thyroid surgery, the initial unilateral approach may be appropriate. However, even in these rare situations it may be possible to expeditiously examine the contralateral side and exclude or reduce the possibility of additional large hyperfunctioning residual glands. Many parathyroid reoperations are straightforward and result in success with unilateral exploration on the side of the missing gland. In the older age group with an increased incidence of multiple gland disease, and patients with asymptomatic or normocalcemic primary HPT (with occasional very subtle differences between diseased and normal glands), it is prudent to explore both sides. When multiple gland disease is anticipated such as in initial cases of multiple endocrine neo-

Randall D. Gaz, M.D. Department of Surgery, Harvard Medical School, Boston, Massachusetts, U.S.A. The role of unilateral neck exploration in the initial management of primary hyperparathyroidism (HPT) remains controversial. Many outstanding parathyroid surgeons have claimed good results using the unilateral approach. The high incidence (>80%) of a single gland disease etiology for primary HPT is such that simple removal of the parathyroid adenoma will result in cure of the disease regardless of how many other glands are identified and examined. Pre-operative localization tests may give a target guiding the surgeon to the appropriate side in over 80-90% of cases. A unilateral exploration may be less likely to result in inappropriate removal or damage of additional contralateral normal parathyroids with resultant hypoparathyroidism. It may shorten the duration of surgery and hospitalization. Surgeons who choose to perform unilateral explorations should clearly discuss the increased risk of missing hyperfunctioning contralateral glands with the patient. This includes the understanding that a second operation, albeit in an undissected Unscarred field, may be necessary. However, pre-operative localization tests often miss the adenoma with single gland disease, overlook an additional enlarged gland(s) with multiple gland disease, or mislead the SUrgeon with false positive findings. They also result in additional expense which may not be beneficial to the patient. Incidentally, it is interesting to note that many Outspoken advocates of the bilateral approach also routinely employ localization procedures prior to primary operations for primary blPT, Unilateral exploration may involve partial or total removal of a small gland which may ultimately have been the most normal and ideally suited for preservation of the parathyroids. The difficulty in making a correct intra-operative distinction between single and multiple gland disease on the basis of examining only two glands raises the specter of error in diagnosis and consequent inadequate removal of diseased glands with resultant persistent or recurrent HPT. The subsequent cost of reinvestigation with more sophisticated localization tests and additional hospitalizations and operations may be Considerable. In skilled hands, bilateral exploration can be Performed safely and in a short period of time without undue morbidity or significant increase in cost. Duh and colleagues have astutely pointed out the potential bias in comparing patients undergoing unilateral exploration With those undergoing bilateral exploration. The latter group