1. No category

Treatment Options for Graves Disease:
A Cost-Effectiveness Analysis
Haejin In, MD, MBA, Elizabeth N Pearce, MD, MSc, Arthur K Wong, MBA, MSc, James F Burgess, PhD,
David B McAneny, MD, FACS, Jennifer E Rosen, MD, FACS
First-line treatment for Graves disease is frequently 18 months of antithyroid medication
(ATM). Controversy exists concerning the next best line of treatment for patients who have
failed to achieve euthyroidism; options include lifelong ATM, radioactive iodine (RAI), or total
thyroidectomy (TT). We aim to determine the most cost-effective option.
STUDY DESIGN: We performed a cost-effectiveness analysis comparing these different strategies. Treatment
efficacy and complication data were derived from a literature review. Costs were examined from
a health-care system perspective using actual Medicare reimbursement rates to an urban university hospital. Outcomes were measured in quality-adjusted life-years (QALY). Costs and
effectiveness were converted to present values; all key variables were subjected to sensitivity
analysis.
RESULTS:
TT was the most cost-effective strategy, resulting in a gain of 1.32 QALYs compared with RAI
(at an additional cost of $9,594) and an incremental cost-effectiveness ratio of $7,240/QALY.
RAI was the least costly option at $23,600 but also provided the least QALY (25.08 QALY).
Once the cost of TT exceeds $19,300, the incremental cost-effectiveness ratio of lifelong ATM
and TT reverse and lifelong ATM becomes the more cost-effective strategy at $15,000/QALY.
CONCLUSIONS: This is the first formal cost-effectiveness study in the US of the optimal treatment for patients
with Graves disease who fail to achieve euthyroidism after 18 months of ATM. Our findings
demonstrate that TT is more cost effective than RAI or lifelong ATM in these patients; this
continues until the cost of TT becomes ⬎ $19,300. (J Am Coll Surg 2009;209:170–179.
© 2009 by the American College of Surgeons)
BACKGROUND:
thalmopathy, or in the presence of thyroid nodules. US
guidelines state that RAI should be the next step.2,3 Yet, in
much of the rest of the world, including Europe, Japan, and
South America, continuation of ATM is the modality of
choice.4 Treatment options seem to be influenced by patient, physician, institutional, and geographical preferences. One reason for this regional discrepancy might be
that previous analyses have suggested that there is no clear
benefit or harm when comparing outcomes among these
therapeutic alternatives.
One method of assessing Graves disease management
options is by using cost-effectiveness analysis for treatment
of Graves disease, which has not been examined in the US
context. We aim to add cost-effectiveness analysis to a decision model to shed light on the optimal next step after
failure of initial ATM. Our objective is to compare the costs
and outcomes of the therapeutic strategies for Graves disease in patients failing to become euthyroid after an 18month course of ATM (Fig. 1). This analysis was used to
identify treatment options that will give patients the best
quality of life relative to the additional cost to the US health
A course of antithyroid medication (ATM) is often used as
the initial treatment for patients with newly diagnosed
Graves disease.1 ATM is typically attempted for 12 to 18
months. This results in remission in about 50% of patients,
with the next step for the others being a choice between
continued ATM, definitive treatment with radioactive iodine (RAI), or total thyroidectomy (TT). There seems to be
no consensus about the next best line of treatment if ATM
fails to induce remission, except for special circumstances
such as children, pregnant women, in the setting of ophDisclosure Information: Nothing to disclose.
Abstract presented at the American College of Surgeons 94th Annual Clinical
Congress, San Francisco, CA, October 2008.
Received November 27, 2008; Revised March 20, 2009; Accepted March 23,
2009.
From the Department of Surgery (In, Wong, McAneny, Rosen) and Department of Endocrinology (Pearce), Boston Medical Center, Boston, MA; and
Veterans Administration Boston Healthcare System, Boston University
School of Public Health, Boston, MA (Burgess).
Correspondence address: Haejin In, MD, Department of Surgery, Boston
Medical Center, 88 East Newton St, #C515, Boston, MA 02118. email:
[email protected]
© 2009 by the American College of Surgeons
Published by Elsevier Inc.
170
ISSN 1072-7515/09/$36.00
doi:10.1016/j.jamcollsurg.2009.03.025
Vol. 209, No. 2, August 2009
In et al
Abbreviations and Acronyms
ATM
QALY
RAI
TT
⫽
⫽
⫽
⫽
antithyroid medication
quality-adjusted life-years
radioactive iodine
total thyroidectomy
care system. Throughout our analysis, we took a healthcare system perspective to build our reference patient.
The reference patient was derived from the most common subset of patients with Graves disease. Data estimates
of treatment efficacies and complications were derived
from a thorough literature review (Table 1). Cost measures
were obtained from actual Medicare reimbursements to a
large urban university hospital (Tables 2, 3). Outcomes
were measured in quality-adjusted life-years (QALY). Costs
and effectiveness were both converted to present values.
Sensitivity analysis was used to examine the robustness of
the model to changes in relevant parameters.
METHODS
Case definition
The decision tree model used for analysis adhered to the
reference-patient scenario recommended by the Panel on
Cost-Effectiveness Analysis of Treatment for Graves Disease
171
Cost-Effectiveness in Health and Medicine.5,6 The most
common scenario was used. Confounders were taken out
of the reference patient by limiting it to a narrow subset of
patients. Graves disease is more common in the female
population and most commonly develops between the second and fourth decades of life.7 Our reference patient is a
30-year-old nonpregnant woman without large goiter,
ophthalmopathy, or palpable nodules who has failed to
remain euthyroid after completion of 18 months of ATM.
Decision model
A decision tree model was used to examine the cost effectiveness of the three treatment options. Decision software
(TreeAge Pro 2008 Healthcare; TreeAge Software) was used
in the construction of our decision model. Each treatment
option incorporated all associated events, their probabilities, and costs, including medications; laboratory tests;
clinic visits; treatment costs; costs associated with adverse
events, such as agranulocytosis and recurrent laryngeal
nerve damage; and costs associated with change in treatment for failed intervention. All major side effects were
included in the analysis. Minor side effects that did not lead
to additional cost or major changes in treatment, such as
development of rash (urticaria) or arthralgia, were ex-
Figure 1. Decision analysis of the management of Graves disease. The boxed areas indicate our focus of analysis.
“Fail” and “succeed” refer to the induction of initial remission after 18 months of antithyroid medication.
172
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
J Am Coll Surg
Table 1. Probabilities and Quality-Life Adjustment
Life expectancy
quality adjustment
factor
Variable
Probability for
reference patient (%)
Short-term
Antithyroid medication
Agranulocytosis‡
0.3
Hepatotoxicity
0.2
Radioactive iodine
Failure of first dose§
15.0
Failure of second dose§
18.0
Failure of third dose§
6.5
Operation (total thyroidectomy)
Hematoma requiring reoperation储
0.8
Transient dysphonia (needing
bronchoscopy)储
5.5
Single vocal cord paralysis (need vocal cord
training)储
0.4
Bilateral vocal cord paralysis (requiring
tracheostomy/ICU/vocal cord
medianization)储
0.0
Hypothyroidism
100.0
Hypoparathyroidism (initial
hypocalcemia)储
10.0
Reoperation for recurrent Graves disease
0.0
Longterm
Antithyroid medication
Euthyroid on medications (continuous
need for medication)
Becomes hypothyroid
2–3 every y
Agranulocytosis‡
0.25
Graves not controlled with medications
(need for other treatment, eg,
development of ophthalmopathy,
uncontrolled Graves, goiter, nodule)¶
3.5
Missed thyroid cancer#
0.10
Radioactive iodine
Longterm hypothyroidism
90 plus 2–3 every y
Missed thyroid cancer/RAI-induced
cancer#**
Operation (total thyroidectomy)
(Permanent) hypothyroidism
(Permanent) hypoparathyroidism
0.10
100.0
1.0
Range of
probabilities (%)
Supporting references
for probability*
0.1–0.5
0.1–0.2
11,14,15
11,12,19
10–38
10–20
6.5
11,18,20,21,22
11,22
22
Longterm
(y)
Short-term
(d)†
Supporting
references for
QALY*
39
39
39
39,41
39,41
39,41
39
39
0.8
26
30
30
7
95
95
95
15
0.25
5.5
26
0.25
39
0.25
39
0.3–0.9
28,31,34
0.957
0.0–0.4
100.0
26,28,31
33
0.979
0.99
3.7–10
0.0
26,28
33
0.95
41
0.95
0.99
0.1–0.5
32
11,22,29,30,32,35
11,14,15
30
39,41
39
39,41
3.5
0.0–5.1
32
16,17,18,42,43
0.9
9
30
39
39,41
11,22,31,36
0.99
0.0–5.1
16,17,18,42,43
0.9
100.0
0.6–1.9
33
26,27,31
0.99
0.95
92.8
90 plus 2–3 per
y or 50 at 5 y
and 70 at 10 y
30
39
39
39
30
39,41
39
40
*Reference numbers as from reference list; 1 year is calculated as 365.25 for cost calculations and 365 for quality-adjusted life-years (QALY) calculations.
†
Calculated in days lost (days lost/365).
‡
Of the lifetime complication risk of agranulocytosis, one-third mostly occur soon after drugs are first initiated, and the remaining risk is during the remaining
lifetime. In the calculations, one-third of the 0.25% complication rate was calculated as complication in year 1, and remaining risk was distributed during the
remaining lifetime.
§
Interval between treatments is 2 to 3 months. All 3 reattempts at radioactive iodine (RAI) were calculated to be year-1 events. After failure of third dose, proceed
to operation or longterm medication.
储
All complications related to operations are evident immediately after operation and calculated with initial cost.
¶
Calculated as occurrence in year 10.
#
All cancers were calculated as occurrence in year 20.
**Missed cancer and RAI-induced cancers were calculated together, because both probabilities were extremely small and many times difficult to distinguish cause.
Vol. 209, No. 2, August 2009
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
173
Table 2. Cost of Specific Events (Medicare 2007)
Variable
Routine visit
Office visit
Laboratory tests ( TSH, free T4, total T3, total T4)
Agranulocytosis
Hospital cost
Antibiotics‡
Laboratory tests‡
Consultants‡
Hepatotoxicity
Laboratory tests (CBC, BMP, LFT)
Ultrasonography
Office visits
Thyroidectomy
Operation consultation (initial outpatient visit)
Preoperative laboratory tests (CBC, BMP, T&S)
Operation (total thyroidectomy)
Hospital admission‡
Endocrine consult‡
Laboratory tests (Ca, Mg, Phos)‡
Postoperative visit***
Vocal cord palsy
Transient
ENT consult
Laryngoscopy
Followup office visit
Permanent single vocal cord paralysis
ENT consult
Laryngoscopy
Followup office visit
Voice reeducation treatment
Permanent bilateral vocal cord paralysis
ENT consult
Laryngoscopy
Emergency tracheostomy
Ventilator support‡
ICU admission‡
Vocal cord medianization
Followup office visit
Voice reeducation treatment
Hypoparathyroid
Office visit
Laboratory tests (TSH, free T4, total T3, total T4)
Laboratory tests (calcium, PTH level)
Radioactive iodine
Pre-RAI laboratory tests (␤-HCG, repeat thyroid function tests)
RAIU
RAI treatment
Office visits (same as above but different schedule)
Subtotal cost ($)
Cost category*
Total cost ($)
114.75
49.82
64.93
99213
†
5,830.98
5,830.98
DRG 420
193.67
35.29
108.56
49.82
†
76700
99213
8,783.85
72.50
91.64
8,619.71
99244
†
DRG 290
138.17
43.31
51.55
43.31
99241
31505
99212
43.31
51.55
43.31
105.00
99241
31505
99212
92507
243.17
34,508.67
43.31
51.55
32,781.53
99241
31505
DRG 482
1,483.97
43.31
105.00
31571
99212
92507
49.82
64.93
64.87
99213
179.62
†
†
435.48
75.42
95.78
214.46
49.82
†
78000
79005
99213
174
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
J Am Coll Surg
Table 2. Continued
Variable
Thyroid cancer
Thyroid US
FNA with US
Surgical consultation (initial outpatient visit)
Preoperative laboratory tests (CBC, BMP, T&S)
Operation (total thyroidectomy)
Hospital admission‡
Endocrine consult‡
Laboratory tests (Ca, Mg, Phos)‡
Postoperative visit‡
Pre-RAI laboratory tests (␤-HCG, repeat thyroid function tests)
RAIU
RAI treatment
Thyroid nuclear scan
Subtotal cost ($)
Cost category*
108.56
81.46
72.50
91.64
8,619.71
76536
76942
99244
Total cost ($)
9,627.62
75.42
95.78
214.46
268.09
†
DRG 290
†
78000
79005
78015
*Cost category will refer to Current Procedural Terminology (CPT) code unless denoted DRG.
†
See summation of laboratory and test costs as listed in Table 3.
‡
Included in the cost of DRG above.
BMP, basic metabolic panel; Ca, calcium; CBC, complete blood count; FNA, fine-needle aspiration biopsy; LFT, liver function test; Mg, magnesium; Phos,
phosphate; PTH, parathyroid; RAI, radioactive iodine; RAIU, radioactive iodine uptake; T&S, type and screen blood; T4, thyroxine; T3, triiodothyronine; TSH,
thyroid-stimulating hormone.
cluded. Treatment options and their corresponding monitoring strategies were derived from the American Thyroid
Association,2 the American Association of Clinical Endocrinologist guidelines,3 and literature review. The time horizon for the analysis was the patient’s remaining life expectancy and was calculated to be 50 years using data from
2004 Social Security Administration Actuarial life tables.8
Table 3. Cost of Laboratory and Diagnostic Tests (Medicare
2007)
Variable
Complete blood count
Basic metabolic panel
Liver function panel
Hepatic function panel
Type and screen blood
Thyroid-stimulating hormone
Free thyroxine
Total triiodothyronine
Total thyroxine
Assay of thyroid
(triiodothyronine or thyroxine)
␤-HCG
Assay of vitamin D
Calcium
Parathyroid hormone
Venipuncture
ECG
Chest x-ray
Cost
($)
Current Procedural
Terminology
9.04
11.83
11.42
11.42
13.19
23.47
12.60
19.81
9.61
85027
80048
80076
80076
86903
84443
84439
84480
84436
9.04
10.49
41.36
7.20
57.67
3.00
19.61
37.97
84479
84703
82306
82310
83970
36415
93005
71010
(The decision tree is available by request from Dr In at
[email protected].)
Costs, effectiveness, quality of life data, and
cost-effectiveness analysis
Costs were examined from the health-care system perspective (third-party payor perspective) and did not include
societal costs (non-health care costs), such as opportunity
cost, loss of productivity, wages, or transportation cost.
Cost of treatment was based on the reimbursement structure for Medicare as representative of US payors with some
augmentation. Actual 2007 Medicare reimbursements to a
large urban university hospital were used to calculate cost
of treatment (Tables 2, 3). Actual billing data for the corresponding disease entities according to their DRG or Current Procedural Terminology codes were used to identify
and verify use of health-care resources corresponding to the
choices in treatment options. Medication costs were obtained from average US wholesale prices (Table 4).9 Consumer price index was used to compute an inflation rate of
5% for future health-care costs.10 A standard discount rate
of 3% for both cost and effectiveness was used in accordance with the recommendations of the Panel on CostEffectiveness in Health and Medicine.5,6 Treatment effectiveness, probabilities, and quality-adjusted life
expectancy data were derived from a systematic literature search. We reviewed medical literature from 1980
to 2007 using a structured Medline search supplemented by manual search of bibliographies of selected
articles (Table 1).11-43 Lifetime probabilities were con-
Vol. 209, No. 2, August 2009
In et al
Table 4. Average Drug Costs (Red Book 2008)
Drug Name
Methimazole
5-mg tabs
10-mg tabs
Potassium iodide
1 bottle (30 mL)
Levothyroxine
50-␮g tabs
100-␮g tabs
Propylthiouracil
50-mg tabs
Calcium
1,000-mg tabs
500-mg tabs
Calcitriol
25-␮g tabs
Price ($)
0.44
0.77
12.10
0.30
0.34
0.16
0.06
0.07
1.21
verted to yearly probabilities using a reversal of the
Markov process. Quality-adjusted life expectancy was
used to calculate QALY. Future QALY were also discounted at 3%, as recommended by the Panel on CostEffectiveness in Health and Medicine.5,6 Results were
focused on incremental cost-effectiveness ratio to obtain
the most cost-effective treatment option.
Data with a wide variation or uncertainty of probability were subjected to sensitivity analysis to examine
the impact of the uncertainty of variables used in the
analysis on the output of the decision model. Variables
were examined by using extreme values of reported data
or estimated 95% confidence interval levels. Sensitivity
analysis was also applied to key variables that would
influence medical decisions or have substantial medical
consequences.
Resource use/event pathway
Lifelong ATM
Patients who do not become euthyroid after 18 months of
antithyroid medication can be treated with lifelong lowdose ATM. This treatment is effective but requires continuous monitoring for possible side effects and potential conversion to hypothyroidism. Patients will incur ATM costs.
All patients were assumed to be started on methimazole
and changed to propylthiouracil if a change in ATMs was
needed. Possible side effects of ATMs include agranulocytosis and hepatotoxicity. The probabilities of these occurrences and the costs associated with each complication,
including additional office visits, laboratory tests, and hospitalization for agranulocytosis, were taken into account.
Other effects of choosing longterm ATM, including risk of
missed thyroid cancer; need for additional treatment be-
Cost-Effectiveness Analysis of Treatment for Graves Disease
175
cause of uncontrolled Graves disease; and development of
goiter, nodules, or ophthalmopathy, were all taken into
account.
RAI
An evaluation, including laboratory tests and RAI uptake
scan, is done before RAI treatment. Dosage of RAI used for
treatment was not relevant to this analysis, as payment did
not differ, but differences in failure rates of RAI treatment
were examined in the sensitivity analysis. After the first
dose of RAI, the followup schedule is typically every 2
months for up to 6 months, until the patient becomes
euthyroid, hypothyroid, or it is determined that RAI needs
to be repeated. RAI can be repeated up to two additional
times. In a small number of patients, RAI will not be effective, and these patients will need an operation or will remain on lifelong medication for control of their disease.
The possibility of missed thyroid cancer or development of
thyroid cancer was also estimated. Although transient thyrotoxicosis or radiation thyroiditis has been known to occur after RAI, this usually occurs in patients much older
than our reference patient, and there is no or minimal
associated cost, so these conditions were not included in
our analysis. Patients who were euthyroid after RAI were
assessed semiannually until they became hypothyroid, at
which time they were changed to an annual followup
schedule.
TT
Cost of TT includes a consultation visit, laboratory tests,
and bundled costs of operation, including hospital admission associated with the procedure and a followup clinic
visit. Possible complications of operation, including hematoma requiring reoperation, dysphonia, vocal cord paralysis, and hypoparathyroidism, were taken into account.
Transient dysphonia is reported in a greater number of
patients than is vocal cord paralysis, and these patients were
assumed to be evaluated by laryngoscopy. Patients with
true vocal cord paralysis underwent additional voiceretraining treatment. Patients with bilateral vocal cord paralysis were considered to require ICU admission and
emergency tracheostomy, laryngoplasty, and voice reeducation. All patients were placed on levothyroxine for life and
were monitored annually. Patients showing signs of hypoparathyroidism postoperatively were started on calcium
supplements (calcitriol) and seen in 2 to 4 weeks. For permanently hypoparathyroid patients, frequent visits are initially necessary to adjust medications and check calcium
levels, then annual followup visits were required for life
(Appendix 1, available online).
176
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
J Am Coll Surg
Table 5. Reference Patient Analysis
Strategy
RAI
TT
Cost ($)
23,610
33,195
Incremental cost ($)
Effectiveness (QALY)
Incremental effectiveness (QALY)
ICER ($/QALY)
9,594
25.084
26.407
1.329
7,250
The strategy of antithyroid medication is dominated by a blend of RAI and TT. The dominated option (simple or extended) has been excluded.
ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-years; RAI, radioactive iodine treatment; TT, total thyroidectomy.
RESULTS
Reference patient
TT was both the most effective (by having the most QALY)
and cost-effective (by incremental cost-effectiveness ratio)
strategy for uncomplicated Graves disease in the 30-year-old
reference patient who failed initial 18-month treatment with
ATM. RAI was the least costly option at $23,610 in present
value but also provided the least QALY (25.08 QALY). TT
resulted in a gain of 1.32 QALY compared with RAI, at an
additional cost of $9,593, resulting in an incremental cost
effectiveness ratio of $7,250/QALY (Table 5). This remains
well under the $50,000 cut-off for cost effectiveness, as suggested by the Panel on Cost-Effectiveness in Health and Medicine,5,6 and is the most cost-effective option. The continued
medication strategy (ATM) provided a less cost-effective option than TT, with an incremental cost-effectiveness ratio of
$15,697/QALY over RAI.
Sensitivity analysis
The greatest influence on the outcomes of the results is the
cost of TT. When the cost of TT is ⬍ $19,300, TT is the
most cost-effective option through dominance or extended
dominance. When the cost of TT is ⬎ $19,300, continued
medication becomes the most cost-effective choice at
$15,000/QALY (Fig. 2).
All other variables, including cost of medication, cost of
RAI, patient age, probability of cancer with medication or
RAI, success rate of RAI, rate of hypothyroidism with RAI,
and complication rates of TT, did not influence the overall
result, and TT remained the most cost-effective option
(Appendix 2, available online).
DISCUSSION
Although treatment choices vary by geographical location,
antithyroid medication is often used as initial therapy for
uncomplicated Graves disease.1 There seems to be no consensus about the next best line of treatment if antithyroid
medication fails to induce remission.
Current US practice guidelines, including those set by
the American Thyroid Association;2 the American Association of Clinical Endocrinologists;3 and textbooks, including Werner & Ingbar’s The Thyroid: A Fundamental and
Clinical Text,4 list the treatment choices and clearly define
their role in special circumstances, such as pregnant
women, children, in the setting of ophthalmopathy, or in
the presence of thyroid nodules. In the absence of these
Figure 2. Sensitivity analysis for cost of total thyroidectomy (TT). The best strategy is indicated by the solid line.
*When cost of TT is ⬍ $5,500, the TT strategy dominates all other options. **When cost of TT is between $5,500
and $19,300, TT is the better option through extended dominance. ***When cost of TT is ⬎ $19,300, antithyroid
medication (ATM) strategy becomes more cost effective (lower incremental cost-effectiveness ratio [ICER]) and is
the better strategy.
Vol. 209, No. 2, August 2009
In et al
situations, the best treatment option is unclear, but RAI is
used frequently in the US.
Contrary to the US context, clinicians in much of the
rest of the world, including Europe, Japan, and South
America, prefer to use longterm antithyroid therapy as the
treatment of choice.4 It is unclear why there is a discrepancy
between treatments preferences in the US compared with
the rest of the world. One reason might be the US culture
of choosing treatments that are focused on “cure” or efficacy over effectiveness. Another possible reason for avoiding lifelong ATM is fear of severe, potentially lethal, side
effects, such as agranulocytosis and toxic hepatic failure.
Several factors need to be considered when defining optimal treatment choices for Graves disease. Many initial
fears about the complications of the various treatments
have been mitigated. ATM side effects, such as agranulocytosis and toxic hepatic failure, although occasionally lethal, are extremely rare, with most being reversed by discontinuation of the drug. The fear of RAI because of its
implication in cancer development has yet to be substantiated (although this was included in the sensitivity analysis).
Although unacceptably high rates of vocal cord paralysis (at
least 25%) were reported when surgical techniques were
first developed, in experienced hands, thyroid operations
now have minimal complication rates, with vocal cord paralysis rates occurring in ⬍ 1% of TT patients.33 Although
controversy still exists, most surgeons now regard TT as the
treatment of choice over subtotal thyroidectomy for surgical management of Graves disease.44-46 This is because subtotal thyroidectomy has been unable to produce reliable
rates of euthyroidism without risk of recurrent disease, and
these patients also have been found to inevitably end up
hypothyroid during longterm followup. With TT as the
operation of choice, failure rates, recurrence rates, and
complications associated with reoperation have all decreased, and TT is a safe and reliable treatment choice.
It is now understood that Graves disease usually leads to
hypothyroidism, regardless of treatment modality. With ATM
or RAI therapy, frequent monitoring of thyroid function is
required to monitor for clinical or subclinical hypothyroidism, at which time levothyroxine is indicated. When a stable
dose has been established, followup can occur at longer intervals. WithTT, patients become permanently hypothyroid and
are started on thyroid hormone supplements, which they will
require for life. Followup can occur in longer intervals once
dose is stabilized, because they do not need to be monitored
for development of hypothyroidism.
This awareness of inevitable hypothyroidism with
Graves disease has led to a change in the objective of treatment. Previously, iodine 131 doses and the extent of surgical thyroid excision were selected with the goal of achieving
Cost-Effectiveness Analysis of Treatment for Graves Disease
177
euthyroidism. Instead, hypothyroidism is now the goal of
treatment, to ensure that hyperthyroidism does not recur.
This philosophy challenges the clinician to reexamine the
best treatment algorithm and creates the opportunity for
using cost-effectiveness analyses to discern the optimal
choice.
Notably, there were several factors not included in our
analysis that would make TT more favorable. We assumed
that all patients who had bilateral vocal cord paralysis
would need tracheostomy, ventilatory support, and laryngoplasty. This is an overestimation of the amount of treatment that this complication of TT would require, as most
bilateral vocal cord paralysis patients do not need all or
most of these interventions. This has made the cost of TT
greater in our analysis than likely would actually be realized
in practice.
Another factor is that RAI can make the cytologic interpretation of thyroid cancer more difficult, in addition to
possibly playing a role in its development.47,48 This difficulty in the interpretation of cytology for thyroid nodules
creates a greater need for thyroidectomy for diagnosis of
thyroid cancer and can diminish the benefit of initially
avoiding TT for RAI.
Our decision analysis has several limitations. The study
was performed retrospectively and used current literature
to generate probabilities and quality of life adjustment factors. Data were collected from multiple sources, which included different time frames and locations. We attempted
to diminish this influence by supplementing the analysis
with expert opinion and by testing wide sensitivity analysis
ranges to illustrate the stability of our results. In addition,
Medicare data from a large urban university hospital and
payment structure based on DRG were used in this analysis. The formal usefulness of the study is limited geographically to the US in terms of the extent of reimbursement of
medical services and to payment structures that used the
DRG system or systems similar to that of Medicare, although much of these costs would be roughly proportional
across wider international contexts. A wide range of prices
was investigated with sensitivity analysis, but the DRG
structure of payment was maintained throughout the
study. It is difficult to translate this particular type of episode of care bundled payment structure to other payment
methods seen globally, such as capitation or government
budgeting systems. Personal influences, such as risk aversion to operations or RAI and cultural bias, were not considered when comparing treatment options, except
through short-term quality-adjustment factors. Importantly, this analysis was from a health systems perspective
and did not include societal costs from lost work productivity or other sources.
178
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
This study demonstrates that TT is the most costeffective treatment strategy in the US in patients whose
Graves disease has not remitted after 18 months of ATM.
TT can also afford patients a higher quality of life than
other treatment options. The sensitivity analysis of the decision model shows that this conclusion holds true until the
cost of TT reaches $19,300, more than twice the actual
observed Medicare cost, at which time continued lifelong
ATM becomes the most cost-effective choice.
Author Contributions
Study conception and design: In, Rosen
Acquisition of data: In, Pearce
Analysis and interpretation of data: In, Pearce, Wong, Burgess
Drafting of manuscript: In, Wong, Rosen
Critical revision: Pearce, Burgess, McAneny, Rosen
REFERENCES
1. Brent GA. Clinical practice. Graves’ disease. N Engl J Med
2008;358:2594–2605.
2. Singer PA, Cooper DS, Levy EG, et al. Treatment guidelines for
patients with hyperthyroidism and hypothyroidism. Standards
of Care Committee, American Thyroid Association. JAMA
1995;273:808–812.
3. Baskin HJ, Cobin RH, Duick DS, et al. American Association of
Clinical Endocrinologists medical guidelines for clinical practice
for the evaluation and treatment of hyperthyroidism and hypothyroidism. Endocr Pract 2002;8:457–469.
4. Werner SC, Ingbar SH, Braverman LE, et al. Werner & Ingbar’s
the thyroid: a fundamental and clinical text. 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2005.
5. Gold MR. Cost-effectiveness in health and medicine. New York:
Oxford University Press; 1996.
6. Muennig P. Designing and conducting cost-effectiveness analysis in medicine and health care. San Francisco: Jossey-Bass;
2002.
7. Pearce EN. Diagnosis and management of thyrotoxicosis. BMJ
2006;332:1369–1373.
8. Social Security Administration. Annual statistical supplement
period life table 2004. Available at: http://www.ssa.gov/OACT/
STATS/table4c6.html. Accessed March 13, 2009.
9. Drug topics red book. Montvale, NJ: Thomson Healthcare;
2007.
10. US Department of Health and Human Services. Consumer
price index. Consumer price index and average annual percent
change for all items, selected items, and medical care components Health US 2004. Available at: Available at: http://
www.hhs-stat.net/scripts/topic.cfm?id⫽73. Accessed March 13,
2009.
11. Werner SC, Ingbar SH, Braverman LE, et al. Werner and Ingbar’s the thyroid: a fundamental and clinical text. 7th ed. Philadelphia: Lippincott-Raven; 1996.
12. Cooper DS. Antithyroid drugs. N Engl J Med 1984;311:1353–
1362.
13. Cooper DS. The side-effects of antithyroid drugs. Endocrinologist 1999;9:457–476.
J Am Coll Surg
14. Tajiri J, Noguchi S. Antithyroid drug-induced agranulocytosis:
special reference to normal white blood cell count agranulocytosis. Thyroid 2004;14:459–462.
15. Cooper DS, Goldminz D, Levin AA, et al. Agranulocytosis associated with antithyroid drugs. Effects of patient age and drug
dose. Ann Intern Med 1983;98:26–29.
16. Rieger R, Pimpl W, Money S, et al. Hyperthyroidism and concurrent thyroid malignancies. Surgery 1989;106:6–10.
17. Dobyns BM, Sheline GE, Workman JB, et al. Malignant and
benign neoplasms of the thyroid in patients treated for hyperthyroidism: a report of the cooperative thyrotoxicosis therapy
follow-up study. J Clin Endocrinol Metab 1974;38:976–998.
18. Phitayakorn R, McHenry CR. Incidental thyroid carcinoma in
patients with Graves’ disease. Am J Surg 2008;195:292–297;
discussion 297.
19. Torring O, Tallstedt L, Wallin G, et al. Graves’ hyperthyroidism:
treatment with antithyroid drugs, surgery, or radioiodine—a
prospective, randomized study. Thyroid Study Group. J Clin
Endocrinol Metab 1996;81:2986–2993.
20. Allannic H, Fauchet R, Orgiazzi J, et al. Antithyroid drugs and
Graves’ disease: a prospective randomized evaluation of the efficacy of treatment duration. J Clin Endocrinol Metab 1990;70:
675–679.
21. Tamai H, Nakagawa T, Fukino O, et al. Thionamide therapy in
Graves’ disease: relation of relapse rate to duration of therapy.
Ann Intern Med 1980;92:488–490.
22. Berglund J, Christensen SB, Dymling JF, et al. The incidence of
recurrence and hypothyroidism following treatment with antithyroid drugs, surgery or radioiodine in all patients with thyrotoxicosis in Malmo during the period 1970⫺1974. J Intern
Med 1991;229:435–442.
23. Holm LE, Dahlqvist I, Israelsson A, et al. Malignant thyroid
tumors after iodine-131 therapy: a retrospective cohort study.
N Engl J Med 1980;303:188–191.
24. Holm LE, Hall P, Wiklund K, et al. Cancer risk after iodine-131
therapy for hyperthyroidism. J Natl Cancer Inst 1991;83:1072–
1077.
25. Hanson GA, Komorowski RA, Cerletty JM, et al. Thyroid gland
morphology in young adults: normal subjects versus those with
prior low-dose neck irradiation in childhood. Surgery 1983;94:
984–988.
26. Sawyers JL, Martin CE, Byrd BF Jr, et al. Thyroidectomy for
hyperthyroidism. Ann Surg 1972;175:939–947.
27. Menegaux F, Ruprecht T, Chigot JP. The surgical treatment of
Graves’ disease. Surg Gynecol Obstet 1993;176:277–282.
28. Werga-Kjellman P, Zedenius J, Tallstedt L, et al. Surgical treatment of hyperthyroidism: a ten-year experience. Thyroid 2001;
11:187–192.
29. Wood LC, Ingbar SH. Hypothyroidism as a late sequela in patient with Graves’ disease treated with antithyroid agents. J Clin
Invest 1979;64:1429–1436.
30. Irvine WJ, Gray RS, Toft AD, et al. Spectrum of thyroid function in patients remaining in remission after antithyroid drug
therapy for thyrotoxicosis. Lancet 1977;2:179–181.
31. Falk SA. Thyroid disease: endocrinology, surgery, nuclear medicine, and radiotherapy. New York: Raven Press; 1990.
32. Azizi F, Ataie L, Hedayati M, et al. Effect of long-term continuous methimazole treatment of hyperthyroidism: comparison
with radioiodine. Eur J Endocrinol 2005;152:695–701.
33. Weber KJ, Solorzano CC, Lee JK, et al. Thyroidectomy remains
an effective treatment option for Graves’ disease. Am J Surg
2006;191:400–405.
Vol. 209, No. 2, August 2009
In et al
34. Palit TK, Miller CC 3rd, Miltenburg DM. The efficacy of thyroidectomy for Graves’ disease: a meta-analysis. J Surg Res 2000;
90:161–165.
35. Weetman AP. Graves’ disease. N Engl J Med 2000;343:1236–
1248.
36. Orgiazzi J. Management of Graves’ hyperthyroidism. Endocrinol Metab Clin North Am 1987;16:365–389.
37. Miller RD, Cucchiara RF. Anesthesia. Philadelphia: Churchill
Livingstone; 2000.
38. Tengs TO, Wallace A. One thousand health-related quality-oflife estimates. Med Care 2000;38:583–637.
39. Vidal-Trecan GM, Stahl JE, Durand-Zaleski I. Managing toxic
thyroid adenoma: a cost-effectiveness analysis. Eur J Endocrinol
2002;146:283–294.
40. Zanocco K, Angelos P, Sturgeon C. Cost-effectiveness analysis of
parathyroidectomy for asymptomatic primary hyperparathyroidism. Surgery 2006;140:874–881; discussion 881⫺882.
41. Vidal-Trecan GM, Stahl JE, Eckman MH. Radioiodine or
surgery for toxic thyroid adenoma: dissecting an important
decision. A cost-effectiveness analysis. Thyroid 2004;14:
933–945.
Cost-Effectiveness Analysis of Treatment for Graves Disease
179
42. Farbota LM, Calandra DB, Lawrence AM, et al. Thyroid carcinoma in Graves’ disease. Surgery 1985;98:1148–1153.
43. Kraimps JL, Bouin-Pineau MH, Mathonnet M, et al. Multicentre study of thyroid nodules in patients with Graves’ disease. Br J
Surg 2000;87:1111–1113.
44. Lal G, Ituarte P, Kebebew E, et al. Should total thyroidectomy
become the preferred procedure for surgical management of
Graves’ disease? Thyroid 2005;15:569–574.
45. Ku CF, Lo CY, Chan WF, et al. Total thyroidectomy replaces
subtotal thyroidectomy as the preferred surgical treatment for
Graves’ disease. A N Z J Surg 2005;75:528–531.
46. Barakate MS, Agarwal G, Reeve TS, et al. Total thyroidectomy is
now the preferred option for the surgical management of
Graves’ disease. A N Z J Surg 2002;72:321–324.
47. Granter SR, Cibas ES. Cytologic findings in thyroid nodules
after 131I treatment of hyperthyroidism. Am J Clin Pathol
1997;107:20–25.
48. Centeno BA, Szyfelbein WM, Daniels GH, et al. Fine needle
aspiration biopsy of the thyroid gland in patients with prior
Graves’ disease treated with radioactive iodine. Morphologic
findings and potential pitfalls. Acta Cytol 1996;40:1189–
1197.
179.e1
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
J Am Coll Surg
Appendix 1. Followup Schedules
Initiation of ATM
Initial
Then
Then
Continue life-long ATM
Then
Hypothyroidism after ATM treatment
Initial
Then
Then
After RAI
Initial
Then
Then
Then
Repeat RAI needed for treatment failure (up to 3 times)
Initial
Then
Then
Then
Then
Then
Then
Then
After TT
Initial
Initial
Then
Then
HypoPTH (permanent) after TT
Initial
Initial
Then
Then
Then
HypoPTH (transient) after TT
Initial
Initial
Then
Then
After TT for thyroid cancer
Initial
Then
Then
Then
Then
Then
Every 4–6 wk for 3–6 mo
Every 2–3 mo
Every 4–6 mo
Every 6 mo
Every 4–6 wk for 3–6 mo
Every 6 mo
Yearly
Every 4–6 wk for 3 mo
Every 2–3 mo
Every 4–6 mo
Every 6 mo
RAI
F/u every 2 wk until wk 6
RAI
F/u every 2 wk until wk 6
RAI
F/u every 2 wk until wk 6
F/u every 4–6 wk until decide lifelong ATM versus TT
surgery
Initiate medication or surgery
1 visit 2 mo postoperative
Every 4–6 wk for 3–6 mo
Every 6 mo
Yearly
1 visit 2 mo post-operatively
Every 4–6 wk for 3–6 mo
Every 3 mo
Every 6 mo
Yearly (after 2nd y)
1 visit 2 mo postoperative
Every 4–6 wk for 3–6 mo
Every 3 mo
Yearly
At 2 wk
RAI at 6–8 wk
Post-treatment scan 1 wk after RAI
Every 4–6 wk for 3–6 mo
Every 6 mo for 10 y
Yearly (after 10 y)
ATM, antithyroid medications; HypoPTH, hypoparathyroid; RAI, radioactive iodine; TT, total thyroidectomy.
Vol. 209, No. 2, August 2009
In et al
Cost-Effectiveness Analysis of Treatment for Graves Disease
179.e2
Appendix 2. Select Details of the Sensitivity Analysis
Range of variables examined
Cost (in dollars)
ATM
RAI
TT
Age at presentation
Probabilities
ATM
Agranulocytosis
Hepatotoxicity
Rate of hypothyroidism
Uncontrolled Graves disease
Thyroid cancer
RAI
Failure of RAI
Thyroid cancer
TT
Hematoma
Transient dysphonia
Single cord paralysis
Bilateral vocal cord paralysis
Permanent
hypoparathyroidism
Results
100–1000
50–500
5000–50000
20–50
No difference
No difference
Different
No difference
0.001–0.005/year
0.001–0.002/year
0.01–0.03/year
0.001–0.01/year
0.0–0.0021/year
No difference
No difference
No difference
No difference
No difference
0.0–0.5
No difference
0.0–0.0021/year
No difference
0.0–0.05
0.0–0.1
0.001–0.05
0.0–0.05
No difference
No difference
No difference
No difference
0.0–0.1
No difference
Details
See text
The analysis was limited to minimize cofounders
that occur at extremes of age. In this age
range, no difference in outcome was observed.
As the age of presentation became greater,
there was a trend towards lower ICER,
suggesting that TT continues to be the most
cost-effective option even at older ages.
TT continued to be the most effective strategy
even when failure rates of RAI approached
0%. As the rate of failure of RAI decreased,
the incremental cost needed to gain 1 QALY
with the TT strategy became greater, but still
remained well above the threshold of changing
the dominance of the strategy.
Increase in probability caused increase in ICER
but did not change the dominant strategy. As
rates of bilateral vocal cord paralysis reached
1%, the ICER was $7,530/QALY and reached
$8,674/QALY at 5%.
Increase in probability caused an increase in
ICER but did not change the dominant
strategy. As the rate of permanent
hypoparathyroidism after TT reached 10%,
the ICER reached $13,823/QALY.
ATM, antithyroid medications; ICER, incremental cost-effectiveness ratio; QALY, quality adjusted life years; RAI, radioactive iodine; TT, total thyroidectomy.