Cost-effectiveness analysis ofparathyroidectomy for asymptomaticprimary hyperparathyroidismKyle Zanocco, BS, Peter Angelos, MD, PhD, and Cord Sturgeon, MD, Chicago, Ill

Background. Controversy exists concerning the best treatment for asymptomatic primaryhyperparathyroidism (PHPT) when the National Institutes of Health consensus conference criteria forparathyroidectomy are not met. We hypothesized that parathyroidectomy would be more cost-effectivethan observation or pharmacologic therapy for these patients.Methods. Cost-effectiveness analysis was performed comparing treatment strategies for asymptomaticPHPT. Treatment outcomes, their probabilities, and costs were identified on the basis of literature andcost database review. Outcomes were weighted by using established quality-of-life utility factors.Sensitivity analysis was used to examine the uncertainty of costs and utility estimates in the model.Results. The incremental cost-effectiveness ratio for parathyroidectomy was $4778 per quality-adjustedlife year (QALY) gained. Operation remained cost-effective until the average cost of parathyroidectomyincreased from the estimated value of $4778 to $14,650. Pharmacologic therapy was not cost-effectiveunless the annual cost of therapy decreased from an estimated $7406 (for cinacalcet) to $221.Parathyroidectomy ceases to be preferred over monitoring if a quality-of-life difference is notdemonstrable after curative operation.Conclusions. Parathyroidectomy is more cost-effective than observation for managing asymptomaticPHPT patients who do not meet National Institutes of Health criteria for parathyroidectomy.Furthermore, pharmacologic therapies with a greater than $221 annual cost were not cost-effective inthis model. (Surgery 2006;140:874-82.)

From the Department of Surgery, Division of Gastrointestinal & Endocrine Surgery, Northwestern University

Feinberg School of Medicine

Primary hyperparathyroidism (PHPT) occursin an estimated 0.2% to 0.5% of the US popula-tion. There are approximately 100,000 new casesin the United States each year and approximately80 to 85% of cases of PHPT are caused by a singleparathyroid adenoma. Operative removal ofadenomatous or hyperplastic parathyroid tissueis the only definitive cure for PHPT. The successof parathyroidectomy (PTX) as a treatment forPHPT, with or without preoperative localization,is higher than 95% in the hands of experiencedsurgeons. With the development of focused or

Presented at the 27th Annual Meeting of the American Associ-ation of Endocrine Surgeons, New York, New York, May, 2006.

Reprint requests: Cord Sturgeon, MD, Assistant Professor ofSurgery, Northwestern University Feinberg School of Medicine,Division of Gastrointestinal & Endocrine Surgery, 201 E HuronSt, Galter 10-105, Chicago, IL 60611-2908. E-mail: csturgeon@nmff.org.

0039-6060/$ - see front matter

© 2006 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2006.07.032

874 SURGERY

minimally invasive techniques and the improve-ment in preoperative and intraoperative localiza-tion tools, PTX has evolved into a procedurecommonly performed in an outpatient settingat many specialized centers across the UnitedStates.1-3

There is general agreement that PTX is indi-cated for all patients with symptomatic PHPT, orasymptomatic patients who are younger than 50years old or who have serum calcium greater than1 mg/dL above the upper limit of normal.4 A mea-surable benefit from PTX has been demonstratedin even mildly symptomatic patients with respect tobone mineral density, quality of life, and psycho-logical function.5 After PTX, symptoms are im-proved in approximately 75% to 85% of patientswith PHPT, including many of those patients whoare considered “asymptomatic” preoperatively.4,6

Recent National Institutes of Health (NIH) recom-mendations suggest that patients with mild asymp-tomatic PHPT who do not meet the consensuscriteria for PTX can be monitored7; however, con-

troversy exists concerning the most effective treat-

Surgery Zanocco, Angelos, and Sturgeon 875Volume 140, Number 6

ment protocol for asymptomatic PHPT patientswho do not meet NIH criteria for parathyroidec-tomy.

Cost analyses performed before the introductionof focused or minimally invasive PTX probably donot reflect the current cost-effectiveness of oper-ative treatment. Before the last decade, PTXcommonly required a longer operation and aninpatient stay of several days. Despite these re-quirements, operation was found to be less costlythan the medical treatment of PHPT when thetime interval required for medical treatment ofPHPT exceeded 5.5 years.8 With the develop-ment of focused or minimally invasive operativetechniques and pharmaceutical agents availablefor the treatment of hyperparathyroidism, addi-tional analyses are now required to determinethe true cost-effectiveness of the currently avail-able medical and operative treatment alterna-tives.9 We performed cost-effectiveness analysisfrom a third-party-payer perspective by comparingthe operative, observational, and pharmacologictreatment strategies for patients with asymptomaticPHPT who do not meet the current NIH criteriafor parathyroidectomy.

METHODSCase definition. The Reference Case recommen-

dations outlined by the Panel on Cost-Effectivenessin Health and Medicine were followed during theconstruction of a decision tree model that analyzedthe three management alternatives for asymptom-atic PHPT.10 The model’s reference case scenarioconsisted of a 60-year-old asymptomatic PHPT pa-tient who does not meet the current NIH criteriafor parathyroidectomy, but is a healthy candidatefor operation. The time horizon for the analysiswas the patient’s remaining life expectancy,which was calculated to be 22 years on the basisof 2005 data published in the United States So-cial Security Administration Period Life Table. Aliterature review was conducted to obtain esti-mates of the costs, health effects, and probabili-ties used in the model.

Decision model. Decision software (TreeAge Pro2005 Suite; TreeAge Software, Williamstown, Mass)was used in the construction of a decision modelfor the treatment of the reference case.10,11 Thecomplete decision tree is shown in Figure 1. Differ-ent treatment pathways were created for the threealternatives: (1) monitoring the asymptomatic pa-tient (the current practice guideline), (2) prescrib-ing pharmacologic intervention with calcimimetics,

or (3) performing parathyroidectomy. The event

pathways and probabilities used in the model arebased on a literature review of current outcomesresearch concerning the medical and operativemanagement of asymptomatic PHPT in the UnitedStates. The selected probabilities are shown inTable I.

A 10-year follow-up horizon was assumed for themonitoring pathway. According to observationaldata reported by Silverberg et al,12 approximatelyone fourth of patients with asymptomatic PHPTwill develop disease progression after 10 years offollow-up. On the basis of these data, the referencecase scenario assumed a 27% probability of thedevelopment of symptomatic disease for the mon-itoring arm.

The model assumed that the patient is a surgicalcandidate with no prior neck operation and thatPTX could be performed via cervical incision only(ie, no ectopic glands requiring sternotomy, thora-cotomy, or thoracoscopy). Fifty percent of initialparathyroidectomies were assumed to be outpatient(�24-hour admission) procedures.1-3 Cure rates for

Fig 1. Complete decision model for the management ofasymptomatic primary hyperparathyroidism. PTX, Para-thyroidectomy; RLN, recurrent laryngeal nerve.

initial and repeat procedures were assigned probabil-

ealth;

876 Zanocco, Angelos, and Sturgeon SurgeryDecember 2006

ities of 95% and 90%, respectively.13-15 Relevant long-term complications of PTX included permanentrecurrent laryngeal nerve (RLN) damage and per-manent hypoparathyroidism. The risk of long-termcomplications was set at 1% for the first PTX and5% for repeat operation.16 These assumptionsshould overestimate permanent complication rateand underestimate cure rate in the hands of expe-rienced surgeons. The risks of temporary RLNdamage or temporary hypoparathyroidism werenot factored into the model.

Costs. Direct medical costs of the monitoring,pharmacologic intervention, and PTX treatmentstrategies were estimated by using reported Medi-care charge and reimbursement data, and are re-ported in Table II in 2005 US dollars.10,11 Coststhat were outside the health care system (eg, trans-portation and lost-productivity costs) were not in-cluded in the analysis. Only costs that differedamong the 3 treatment strategies were included.10,11

The current practice recommendations for the man-agement of asymptomatic PHPT provided by a panelof experts at a NIH consensus development confer-ence in 2002 were used to determine the health careresources corresponding to the monitoring strategy.7

Current practice recommendations and billingdata from a single tertiary referral hospital wereused to identify health care resources correspond-ing to parathyroidectomy.7 All the health care com-ponents of each treatment strategy were assignedtheir corresponding Diagnosis-Related Groups orCurrent Procedural Terminology (CPT) codes.

We assumed that patients undergoing PTX wouldreceive preoperative ultrasonography and sestamibiscanning, preoperative electrocardiogram, and a sur-

Table I. Summary of probability assumptions used

Assumption

Probability of developing NIH indication for PTXduring follow-up

Probability that first PTX will be outpatient procedure

Cure rate after first PTX

Cure rate after repeat PTXSignificant complication rate after first PTX

Significant complication rate after repeat PTX

Probability of observation of persistent/recurrent diseasafter PTX without complication

HypoPTH, permanent hyperparathyroidism; NIH, National Institutes of H

gical consultation. The costs of performance and in-

terpretation of these tests were included. Costsassociated with the use of different operating roomtechnologies including intraoperative parathyroidhormone monitoring were not itemized, but areincluded in the hospital costs estimate for parathy-roidectomy. This treatment of operating roomcosts is based on the Medicare prospective paymentsystem and maintains the third-party-payer perspec-tive of this analysis.

Inpatient hospital costs were estimated by calcu-lating a Medicare cost-to-charge ratio for Diagnosis-Related Groups 289, Parathyroid Procedures, byusing data from the Medicare Provider Analysis andReview database. The computed ratio of 0.295 wasthen multiplied by the national median inpatientcharge for PTX reported by the Nationwide Inpa-tient Sample.11 Outpatient hospital costs were esti-mated on the basis of median rates reported by theMedicare Hospital Outpatient Prospective PaymentSystem for the corresponding outpatient CPTcodes.11 Cost estimates for physician fees werebased on median Medicare reimbursement ratesfor the corresponding CPT code, and ambulatorylaboratory costs were based on Medicare’s ClinicalLaboratory Fee Schedule. Medication costs werebased on average US wholesale prices.17 Consumerprice index data for health care from 1995 to 2005were used to compute a future health care costinflation rate of 3.9%.11 A discount rate of 3% wasapplied to all future costs in the model.10,11

Effectiveness. Quality-adjusted life years(QALYs) were used as the measure of effective-ness in the model.10,11 Each health outcome wasassigned a quality adjustment factor that was mul-tiplied by the time spent with the corresponding

e decision model

Probability used in the model Supporting references

.27 Silverberg12

.50 Udelsman1

Pellitteri2

Grant3

.95 Carneiro13

Hedback14

.90 al-Fehaily15

.01 (.005 HypoPTH) Fahy16

(.005 RLN damage).05 (.01 HypoPTH) Fahy16

(.04 RLN damage).06 Hedback14

PTX, parathyroidectomy; RLN, recurrent laryngeal nerve.

in th

e

outcome to produce the total QALYs realized by

ECG, el

Surgery Zanocco, Angelos, and Sturgeon 877Volume 140, Number 6

each pathway in the decision tree.11,18 FutureQALYs were discounted at 3% as recommended bythe Panel on Cost-Effectiveness in Health and Med-icine.10 Quality adjustment factors used in themodel are listed in Table III.

Patients in the model who were cured by suc-cessful PTX were assumed to be in perfect health(for comparison purposes) and were assigned aquality adjustment factor of 1. Calcimimetic treat-ment has been shown to be effective in establishingnormocalcemia in select patients with PHPT.19 Forthis reason, asymptomatic patients successfully

Table II. Costs included in the decision model

Monitoring parameter

Serum calciumSerum creatinineBone density, lumbar spineBone density, forearm and wristFollow-up appointmentsTotal per year

Pharmacologic therapy Dosage

Cinacalcet 30 mg bid19 $Alendronate 40 mg/day $Alendronate 75 mg 1� weekly $Risedronate 5 mg/day $Risedronate 35 mg 1� weekly $Calcitonin (salmon) 200 IU (.09 mL)/day $

PTX CPT code

Hospital costsPTX (DRG 289) 60500Ultrasound 76536ECG 93005Parathyroid Imaging 78070Sestamibi dose A9500

Physician costsAnesthesiology 00320Surgery office consult 99242Surgery PTX 60500Radiology 76536-26Cardiology 93010Nuclear medicine 78070-26Pathology 88305-26

TotalEstimated treatment costs for

surgical complicationsVocal cord medialization

after RLN damage$8479.70

Annual cost of calcitrioltreatment for long-termhypoparathyroidism

$705.94

CPT, Current Procedural Terminology; DRG, Diagnosis-Related Groups;

treated with calcimimetics were also assigned a

quality adjustment factor of 1. Multiple studies sug-gest that asymptomatic PHPT patients report a di-minished quality of life that is improved withoperation.4,6 Estimates of the quality adjustmentfactors for asymptomatic PHPT and for the compli-cations of PTX were derived from a systematic Med-line literature search.20-23 Published SF-36 HealthSurvey results were identified that includedscores before and after PTX in asymptomaticPHPT patients as well as scores after vocal cordmedialization after RLN injury. Statistically sig-nificant SF-36 dimensions were then converted to

CPT code Cost

82310 $19.4382565 $9.6776075 $136.0776076 $41.4699213 $103.84

$310.47

holesale price(quantity) Estimated daily cost

Total costper year

(30 mg 30.0’s) $19.44 $7095.60(40 mg 30’s) $6.21 $2266.77

(75 mg 4’s) $3.02 $1102.30(5 mg 30’s) $2.64 $962.14(35 mg 4’s) $2.64 $962.30(3.7 mL) $2.31 $844.51

tient PTX Outpatient PTX Repeat PTX

3923.07 $2103.10 $3923.0781.19 $81.19 $81.1922.42 $22.42 $22.42170.87 $170.87 $170.8769.09 $69.09 $69.09

106.56 $106.56 $106.5691.36 $91.36 $91.36966.68 $966.68 $1123.4329.06 $29.06 $29.069.05 $9.05 $9.0542.96 $42.96 $42.9641.83 $41.83 $41.835554.14 $3734.17 $5710.89

ectrocardiogram; PTX, parathyroidectomy.

W

291.60186.3184.5679.0873.8295.12

Inpa

$$$$$

$$$$$$$$

quality-of-life adjustment factors through the use

878 Zanocco, Angelos, and Sturgeon SurgeryDecember 2006

of a multivariate regression model developed byNichol et al.24

Sensitivity analysis. The reference case scenariowas tested by using sensitivity analysis to identify theuncertainty level of the results.10,11,18 Each variablein the model was tested independently across arange of possible values to determine the impact ofdifferent assumptions on the cost-effectiveness re-sults. Key assumptions were tested with the use ofmultivariate sensitivity analysis, in which the effectsof simultaneously changing multiple variables wereanalyzed. Key assumptions included the costs ofPTX, calcimimetic therapy, and monitoring asymp-tomatic PHPT; the complication rates of PTX; andthe quality adjustment factors for RLN damage andasymptomatic PHPT.

RESULTSReference case. The monitoring strategy was the

least costly as well as the least effective with anexpected cost of $4209 for a discounted effective-ness of 15.766 QALYs. The incremental cost-effec-tiveness ratio (ICER) for PTX was $4778 per QALYgained over the monitoring strategy, reflecting again of 0.163 QALYs at an additional cost of $777.When compared with PTX, the calcimimetic treat-ment option produced a gain of 0.008 QALY at anadditional cost of $176,097, resulting in an ICER of$20,995,772 per QALY gained. These referencecase results are summarized in Table IV.

Interventions that have ICERs exceeding a thresh-old of $50,000 per QALY gained are not typicallyconsidered cost-effective.10,11 Applying this thresholdto the reference case reveals that PTX is cost-effective,

Table III. Quality adjustment factors used in the d

Outcome

Stable asymptomatic PHPT

Cured with RLN damage requiring vocal cord medializa

Long-term hypoparathyroidismStable asymptomatic disease with RLN damage

Symptomatic PHPT

Symptomatic PHPT with RLN damage

PHPT, Primary hyperparathyroidism; RLN, recurrent laryngeal nerve.

but calcimimetic therapy is not cost-effective.

Sensitivity analysis. As the estimated annual costof monitoring approached $0, the ICER for theoperative management strategy increased butremained under $25,000 per QALY. The operativestrategy dominated the monitoring strategy whenthe cost of monitoring was greater than $397 peryear: Monitoring costs exceeding $397 per yearmade PTX both more effective and less costly thanmonitoring (Fig 2). Operation also dominatedthe monitoring strategy when the average cost ofPTX was less than $3400. Operation remained cost-effective until the cost of PTX increased to $14,650(Fig 3). The calcimimetic treatment ICER re-mained above the $50,000 per QALY thresholduntil the annual cost of treatment decreased from$7406 per year to $221 per year. When the cost oftreatment was between $221 per year and $172 peryear, the theoretic gain in quality of life with calci-mimetic treatment became cost-effective. Pharma-cologic treatment became the dominant (ie, mosteffective and least costly) strategy when the totalcost of treatment fell below $172 per year (Fig 4).

Outcomes from the decision model were af-fected by changes in the estimated quality adjust-ment factors for stable asymptomatic disease andRLN damage. Figure 5 demonstrates that PTX re-mained the preferred strategy until the asymptom-atic disease quality adjustment factor exceeded0.998. Multivariate sensitivity analysis of combinedcomplication and quality adjustment factor rangesfor RLN damage is illustrated in Figure 6. As thequality adjustment factor for RLN damage de-creased, the probability of RLN injury necessaryto achieve or maintain cost-effectiveness also

n model

Life expectancy qualityadjustment factor Supporting references

0.987 Burney20

Sheldon21

0.979 Spector22

Vidal-Trecan23

0.950 Vidal-Trecan23

0.957 Burney20

Sheldon21

Spector22

Vidal-Trecan23

0.897 Burney20

Sheldon21

0.877 Burney20

Sheldon21

Spector22

ecisio

tion

decreased.

Surgery Zanocco, Angelos, and Sturgeon 879Volume 140, Number 6

DISCUSSIONThis study demonstrates that PTX is a cost-effec-

tive treatment strategy for patients with asymptom-atic PHPT and affords patients a higher qualityof life than a monitoring strategy. The major con-tributors to this result include an increasing fre-quency of cost-saving outpatient procedures andthe conclusions of multiple investigations that dem-onstrate quality-of-life improvement after PTX evenin patients thought to be asymptomatic fromPHPT.1-6 Sensitivity analysis of the decision modelshows that the average cost of PTX can increase to3 times the estimated amount used in this studybefore PTX ceases to be cost-effective. Sensitivityanalysis also shows that PTX is only an effective

Table IV. Reference case results

Strategy Cost Incremental cost Effec

Observation $4209 — 15PTX $4986 $777 15Pharmacologic

treatment(cinacalcet)

$181,083 $176,097 15

ICER, Incremental cost-effectiveness ratio; PTX, parathyroidectomy.*Quality-adjusted life years.†Per quality-adjusted life year.

Fig 2. The incremental cost-effectiveness ratio for oper-ative is shown as a function of the annual cost of obser-vation. The threshold for cost savings is the point whereparathyroidectomy dominates the observation strategy.When the annual cost of observation exceeds this thresh-old value of $397, parathyroidectomy is less costly inaddition to being more effective than observation. QALY,Quality-adjusted life year.

treatment strategy if a difference in quality of life

exists between patients with asymptomatic PHPTand those cured of disease.

Pharmacologic treatment with the calcimimeticagent cinacalcet was shown in one randomized,double-blind, placebo-controlled trial to normalizecalcium in 73% of study subjects with PHPT.19 Thecurrent wholesale price of cinacalcet, however,makes theoretical gains in quality of life withpharmacologic treatment prohibitively expensive(�$20 M per QALY gained). Furthermore, thisanalysis compared PTX and pharmacologic ther-apy with cinacalcet at the drug’s lowest demon-strated effective dose: 30 mg twice daily.19 Themodel assumed that patients undergoing this ther-apy would achieve the same increase in quality of

sIncrementaleffectiveness Cost/effectiveness ICER

— $267† —0.163* $313† $4778†0.008* $11,362† $20,995,772†

Sensitivity Analysis ofAverage Cost of Routine Parathyroidectomy

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

$0 $2,500 $5,000 $7,500 $10,000 $12,500 $15,000 $17,500 $20,000

Average Cost of Routine Parathyroidectomy

Incr

em

en

tal C

ost

/Eff

of S

urg

ery

($

/QA

LY

) Threshold for acceptableICER = $14,650

Reference Case Value = $4,664

Threshold forcost savings = $3,400

Fig 3. The incremental cost-effectiveness ratio (ICER)for operative treatment is shown as a function of theaverage cost of parathyroidectomy for a patient withoutprevious neck surgery. Operation dominates observationwhen the average cost of parathyroidectomy is below$3400. When the average cost exceeds $14,650, the op-erative strategy ceases to be cost-effective by exceedingthe cost-effectiveness threshold ICER of $50,000 perquality-adjusted life year (QALY).

tivenes

.766*

.929*

.937*

life as those patients undergoing successful, com-

880 Zanocco, Angelos, and Sturgeon SurgeryDecember 2006

plication-free operation with 100% efficacy and noadverse events. Even if these assumptions were true,our analysis reveals that cinacalcet would still needto be discounted to less than 5% of its current price

Sensitivity Analysis ofTotal Annual Cost of Pharmacologic Treatment

0

1

10

100

1,000

10,000

100,000

1,000,000

10,000,000

100,000,000

$0 $1,000 $2,000 $3,000 $4,000 $5,000 $6,000 $7,000 $8,000

Total Annual Cost of Pharmacologic Treatment

Inc

rem

en

tal C

os

t/E

ff o

f P

ha

rma

co

log

ic

Tre

atm

en

t ($

/QA

LY

)

Reference Case Value = $7,406Threshold for

acceptableICER = $221

Threshold forcost savings = $172

Fig 4. The incremental cost-effectiveness ratio (ICER)for the pharmacologic treatment strategy is shown as afunction of the total annual cost of pharmacologic treat-ment. Pharmacologic therapy becomes cost-effectiveas the annual cost decreases to $221 per year andbecomes the dominant strategy when the cost is lowerthan $172 per year. QALY, Quality-adjusted life year.

Sensitivity Analysis ofQuality Adjustment Factor fo r Stable Asymptomatic Disease

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

0.98 0.982 0.984 0.986 0.988 0.99 0.992 0.994 0.996 0.998 1

Quality Adjustment Factor for Stable Asymptomatic Disease

Incr

emen

tal C

ost/

Eff o

f Sur

gery

($/Q

ALY

)

Threshold for acceptableICER = 0.998

Reference Case Value = 0.987

Fig 5. The incremental cost-effectiveness ratio (ICER)for operative treatment is shown as a function of thequality adjustment factor for stable asymptomatic dis-ease. When the quality adjustment factor exceeds .998,the operative strategy ceases to be cost-effective by ex-ceeding the cost-effectiveness threshold ICER of $50,000per quality-adjusted life year.

before it would be considered cost-effective in

asymptomatic PHPT patients who are otherwisehealthy candidates for PTX.

The decision model has several limitations. Thequality-of-life adjustment factors were determinedindirectly through the conversion of SF-36 scoreswith the use of a regression model. Quality adjust-ment factors determined directly by the time trade-off method have been determined through a surveyof unaffected individuals.25 Our study’s conclu-sions, however, are unchanged by the use of thesedifferent quality adjustment factors. Patients whodevelop symptomatic disease during monitoringwere assumed to have immediate PTX without anytime spent with symptomatic disease, introducingfavorable bias toward the monitoring strategy.Short-term time losses and minor operative compli-cations as well as extremely low probability out-comes including anesthesia-related mortality werenot addressed by the model. All patients were as-sumed to be candidates for PTX and to have hadno prior neck operation. Ectopic glands not ame-nable to removal by cervical incision were not ac-counted for. These omissions introduced favorablebias toward PTX. An outpatient procedure rate of50% was assumed for patients undergoing PTX on

Two-Way Sensitivity Analysis of RLN Complication Rate and Quality Adjustment Factor for RLN Damage

0

0.03

0.06

0.09

0.12

0.15

0.75 0.8 0.85 0.9 0.95 1

Quality Adjustment Factor for RLN DamageFollowing Vocal Cord Medialization

Pro

ba

bili

ty o

f R

LN

Da

ma

ge

Reference Case Value= 0.979, 0.005

MONITORING IS FAVORABLE ON THIS SIDE OF THE THRESHOLD LINE

SURGERY IS FAVORABLE ON THIS SIDE OF THE THRESHOLD LINE

Fig 6. This multivariate sensitivity analysis shows thecombined effects of changes in the probability of recur-rent laryngeal nerve (RLN) damage and the quality ad-justment factor for RLN damage after corrective vocalcord medialization. The threshold line represents com-binations of these values that produce an incrementalcost-effectiveness ratio (ICER) for the operative strategyof $50,000 per quality-adjusted life year (QALY). Thearea to the right of the curve represents values that arebelow this threshold, making parathyroidectomy the fa-vorable strategy. The area to the left of the curve repre-sents values that produce an ICER that exceeds $50,000per QALY, making monitoring the favorable strategy.

the basis of our review of the available literature.

Surgery Zanocco, Angelos, and Sturgeon 881Volume 140, Number 6

This estimate, however, is probably lower than cur-rent outpatient trends. Increasing the outpatientprocedure rate would increase cost savings andfavor PTX. Favorable bias was also introduced forcalcimimetic therapy because the model assumed100% efficacy at the lowest described dose and anabsence of adverse events. Cinacalcet has not beenshown to improve bone density in patients withosteopenia or osteoporosis, and the extra costs oftreating these complications were not includedin the pharmacologic treatment strategy. Theselimitations are obviated by the finding that an inputof a wide range of acceptable values for qualityadjustment factors and costs was shown by sensitiv-ity analysis to produce the same qualitative results.

The NIH consensus development conference onasymptomatic PHPT concluded that patients withmild asymptomatic PHPT who do not meet thecriteria for operative treatment can be monitoredwithout the need for operative intervention.7 Inthis study we found, however, that a cost-effectiveimprovement in quality of life can be achievedthrough PTX in asymptomatic patients who do notmeet the current NIH criteria for parathyroidec-tomy. When compared with a medical monitoringstrategy or cinacalcet therapy, PTX in the hands ofexperienced surgeons provides the optimal treat-ment option for patients with mild asymptomaticPHPT.

The authors wish to thank Amy Boyle and LawrenceMoy for their assistance in interpreting Medicare codingand reimbursement data.

REFERENCES1. Udelsman R. Six hundred fifty-six consecutive explorations

for primary hyperparathyroidism. Ann Surg 2002;235(5):665-70; discussion 670-2.

2. Pellitteri PK. Directed parathyroid exploration: evolutionand evaluation of this approach in a single-institution reviewof 346 patients. Laryngoscope 2003;113(11):1857-69.

3. Grant CS, Thompson G, Farley D, van Heerden J. Primaryhyperparathyroidism surgical management since the introduc-tion of minimally invasive parathyroidectomy: Mayo Clinic ex-perience. Arch Surg 2005;140(5):472-8; discussion 478-9.

4. Eigelberger MS, Cheah WK, Ituarte PH, Streja L, Duh QY,Clark OH. The NIH criteria for parathyroidectomy inasymptomatic primary hyperparathyroidism: are they toolimited? Ann Surg 2004;239(4):528-35.

5. Rao DS, Phillips ER, Divine GW, Talpos GB. Randomizedcontrolled clinical trial of surgery versus no surgery in pa-tients with mild asymptomatic primary hyperparathyroid-ism. J Clin Endocrinol Metab 2004;89(11):5415-22.

6. Sywak MS, Knowlton ST, Pasieka JL, Parsons LL, Jones J. Dothe National Institutes of Health consensus guidelines forparathyroidectomy predict symptom severity and surgicaloutcome in patients with primary hyperparathyroidism?Surgery 2002;132(6):1013-9; discussion 1019-20.

7. Bilezikian JP, Potts JT Jr, Fuleihan Gel H, et al. Summary

statement from a workshop on asymptomatic primary hy-

perparathyroidism: a perspective for the 21st century. J ClinEndocrinol Metab 2002;87(12):5353-61.

8. Heath H 3rd, Hodgson SF, Kennedy MA. Primary hyperpara-thyroidism. Incidence, morbidity, and potential economic im-pact in a community. N Engl J Med 1980;302(4):189-93.

9. Mowschenson PM, Hodin RA. Outpatient thyroid and para-thyroid surgery: a prospective study of feasibility, safety, andcosts. Surgery 1995;118(6):1051-3; discussion 1053-4.

10. Gold MR. Cost-effectiveness in health and medicine. NewYork: Oxford University Press; 1996.

11. Muennig P, Khan K. Designing and conducting cost-effec-tiveness analyses in medicine and health care. 1st ed. SanFrancisco: Jossey-Bass; 2002.

12. Silverberg SJ, Shane E, Jacobs TP, Siris E, Bilezikian JP. A10-year prospective study of primary hyperparathyroidismwith or without parathyroid surgery. N Engl J Med1999;341(17):1249-55.

13. Carneiro DM, Solorzano CC, Irvin GL 3rd. Recurrent dis-ease after limited parathyroidectomy for sporadic primaryhyperparathyroidism. J Am Coll Surg 2004;199(6):849-53;discussion 853-5.

14. Hedback G, Oden A. Persistent disease after surgery forprimary hyperparathyroidism: the long-term outcome. Eur JEndocrinol 2004;150(1):19-25.

15. al-Fehaily M, Clark OH. Persistent or recurrent primaryhyperparathyroidism. Ann Ital Chir 2003;74(4):423-34.

16. Fahy BN, Bold RJ, Beckett L, Schneider PD. Modern para-thyroid surgery: a cost-benefit analysis of localizing strate-gies. Arch Surg 2002;137(8):917-22; discussion 922-3.

17. Red Book. Montvale (NJ): Thomson Healthcare; 2005.18. Petitti DB. Meta-analysis, decision analysis, and cost-effec-

tiveness analysis: methods for quantitative synthesis in med-icine. 2nd ed. New York: Oxford University Press; 2000.

19. Peacock M, Bilezikian JP, Klassen PS, Guo MD, Turner SA,Shoback D. Cinacalcet hydrochloride maintains long-termnormocalcemia in patients with primary hyperparathyroid-ism. J Clin Endocrinol Metab 2005;90(1):135-41.

20. Burney RE, Jones KR, Christy B, Thompson NW. Healthstatus improvement after surgical correction of primary hy-perparathyroidism in patients with high and low preopera-tive calcium levels. Surgery 1999;125(6):608-14.

21. Sheldon DG, Lee FT, Neil NJ, Ryan JA Jr. Surgical treatmentof hyperparathyroidism improves health-related quality oflife. Arch Surg 2002;137(9):1022-6; discussion 1026-8.

22. Spector BC, Netterville JL, Billante C, Clary J, Reinisch L,Smith TL. Quality-of-life assessment in patients with unilat-eral vocal cord paralysis. Otolaryngol Head Neck Surg2001;125(3):176-82.

23. Vidal-Trecan GM, Stahl JE, Eckman MH. Radioiodine or sur-gery for toxic thyroid adenoma: dissecting an important deci-sion. A cost-effectiveness analysis. Thyroid 2004;14(11):933-45.

24. Nichol MB, Sengupta N, Globe DR. Evaluating quality-ad-justed life years: estimation of the health utility index (HUI2)from the SF-36. Med Decis Making 2001;21(2):105-12.

25. Sejean K, Calmus S, Durand-Zaleski I, et al. Surgery versusmedical follow-up in patients with asymptomatic primaryhyperparathyroidism: a decision analysis. Eur J Endocrinol2005;153(6):915-27.

DISCUSSIONDr John A. Chabot (New York, NY). Am I cor-

rect to conclude that the younger a patient is themore favorable the cost analysis would be for the

one-time surgical intervention?

882 Zanocco, Angelos, and Sturgeon SurgeryDecember 2006

Mr Kyle Zanocco. Absolutely, you are correct.Because the time frame for monitoring the patientswould be lengthened and it would be more costlybecause you would be adding the costs as life goes on.

Dr James Norman (Tampa, Fla). I think this wasa wonderful study and I really enjoyed the presen-tation. My only complaint is that you didn’t givethis talk in front of a bunch of endocrinologists,because I think that this would be a wonderfulpaper in an endocrinology journal.

I understand that some of the most recent datashows that using Calcinicet for primary hyperpara-thyroidism may not provide protective effects onbone loss and those patients still continue to getworsening of their bones and worsening of osteo-porosis. So I am not sure that that drug is going toplay much of a role in primary hyperthyroidism.

My question to you is regarding asymptomatichyperparathyroidism. I don’t think that asymptom-atic hyperparathyroidism exists near as often assome people say it does, and I wanted your opinionon that. How often do you think there really isasymptomatic hyperparathyroidism and do youshare my bias that if you ask the right questionsalmost everybody has symptoms.

Mr Kyle Zanocco. From review of the literature Ido agree that it appears that many patients havesymptoms. Unfortunately, I am a medical studentand I—

Dr John A. Chabot. Probably an unfair questionfor a second-year medical student.

Mr Kyle Zanocco. I don’t have much experiencediagnosing it. The answer is yes, though.

Dr Lawrence A. Danto (Truckee, Calif). I toothought it was an excellent paper. It really doescompare some significant measures in cost-effec-tiveness that I haven’t come across in the past thatwere well done. My question to you is this: Have youused this data set to compare the cost-effectivenessof minimally invasive parathyroid surgery versusunilateral parathyroid surgery versus the standardbilateral parathyroid exploration?

Mr Kyle Zanocco. The way we performed ouranalysis was based on the way Medicare pays forthese procedures. But we were able to break itdown based on inpatient and outpatient initial op-erations, so that would reflect probably a differencebetween minimally invasive and unilateral versusstandard procedures. But we did not explicitly cost

each of those procedures.

Dr. Nancy D. Perrier (Houston, Tex). Great pre-sentation and great data that I think is going to beinteresting for all of us. My only question or issue ison your method slide looking at the cost, which iswhat was just brought up, whether we are talkingabout billing or collections, the actual cost. Butunder there, if you look at the parathyroid imagingat $170 and the sestamibi dose at $69, I think themajority of us in the room that do this routinelyand look at these costs know that that is a very lowestimate compared to what we see.

I know that things are a lot bigger and moreexpensive in Texas. But in Texas, the cost of aSPECT/CT, and as we just discussed in a presenta-tion a few presentations ago, the cost of aSPECT/CT in Texas is $3,600. And for those of uswho do minimally invasive parathyroidectomy, thatis one of the common tests we give.

So there really is a difference. And I just want tocaution us, because when we present this datato the endocrinologists, and I think they are goingto want to hear it, we do need to make sure that weback that up, because that slide will be scrutinizedclosely. I don’t know if you have that data, thoseanswers, but some of your sponsors might.

Mr Kyle Zanocco. I would like to comment onthat. We used Medicare reimbursement rates forthese things, the median rate of reimbursement.And I think probably what you are referring to arecharges for patients. And general procedure forcost-effectiveness analysis requires using things thatreflect costs as opposed to charges. That is whythese numbers seem so low. Also, additionally, if welook at our sensitivity analysis, we see that we havea pretty large area to work with where it is stillcost-effective to do parathyroidectomy before wereach unacceptable costs.

Professor Antonio Stiger-Serra (Barcelona,Spain). My question deals with how many of youreally follow the guidelines? Because I come from acountry in which guidelines are just a suggestion.We would operate on many patients who don’t fitthe guidelines. Imagine a patient of 55 with a Tscore of minus 2.3 and a calcium clearance of 300score who doesn’t meet any of the criteria but willgo on to have parathyroidectomy. So I think theissue is more common sense for those unfit forsurgery, but surgery should certainly be done even

for many, many patients without the guidelines.