A cost-effectiveness analysis of adrenalectomy for nonfunctional adrenal incidentalomas: Is there a size threshold for resection?
Tracy S. Wang, MD, MPH,a Kevin Cheung, MSc, MD,b Sanziana A. Roman, MD, and Julie A. Sosa, MD, MA, Milwaukee, WI, Hamilton, Canada, and New Haven, CT
Background. Adrenocortical carcinoma (ACC) is a rare, but aggressive, malignancy. Current American Association of Clinical Endocrinologists (AACE)/American Association of Endocrine Surgeons (AAES) guidelines recommend resection of nonfunctional adrenal neoplasms ≥4 cm. This study evaluates the cost-effectiveness of this approach.
Methods. A decision tree was constructed for patients with a nonfunctional, 4-cm adrenal incident- aloma with no radiographic suspicion for ACC. Patients were randomized to adrenalectomy, surveillance per AACE/AAES guidelines, or no follow-up (“sign-off”). Incremental cost-effectiveness ratio (ICER) includes health care costs, including missed ACC. ICER (dollar/life-year-saved [LYS]) was determined from the societal perspective. Sensitivity analyses were performed.
Results. In the base-case analysis, assuming a 2.0% probability of ACC for a 4-cm tumor, surgery was more cost-effective than surveillance (ICER 35/LYS and $8/LYS, respectively). Sensitivity analysis demonstrated that the model was sensitive to patient age, tumor size, probability of ACC, mortality of ACC, and cost of hospitalization. The results of the model were stable across different cost and complications related to adrenalectomy, regardless of operative approach.
Conclusion. In our model, adrenalectomy was cost-effective for neoplasms >4 cm and in patients <65 years, primarily owing to the aggressiveness of ACC. Current AACE/AAES guideline recommendations for the resection of adrenal incidentalomas ≥4 cm seem to be cost-effective. (Surgery 2012;152:1125-32.)
From the Department of Surgery,” Medical College of Wisconsin, Milwaukee, WI; the Department of Surgery,b McMaster University, Hamilton, Canada; and the Department of Surgery,“ Yale University School of Medi- cine, New Haven, CT
ADRENAL MASSES are detected in approximately 4-6% of patients undergoing abdominal imaging studies for indications other than the suspicion of adrenal disease; the prevalence increases with age and may be found in up to 7% of patients >70 years old.1,2 There is consensus that the presence of an inciden- tally detected adrenal lesion (“incidentaloma”) should prompt concern for functionality and/or malignancy. All incidental adrenal tumors should, therefore, undergo biochemical evaluation for au- tonomous cortisol secretion (hypercortisolism),
Accepted for publication August 10, 2012.
Reprint requests: Tracy S. Wang, MD, MPH, Department of Sur- gery, Medical College of Wisconsin, 9200 W. Wisconsin Avenue, Milwaukee, WI 53226. E-mail: tswang@mcw.edu.
0039-6060/$ - see front matter
http://dx.doi.org/10.1016/j.surg.2012.08.011
hyperaldosteronism, and pheochromocytomas; ad- renalectomy should be performed for all functional tumors as long as there are no contraindications to surgery.36 Tumors that do not meet criteria for resec- tion should undergo hormonal evaluation annually for 5 years, because some tumors may become func- tional over time. 4-10
Adrenocortical carcinoma (ACC) has been re- ported in ≤9% of adrenal incidentalomas.11-14 Imaging studies should be evaluated carefully to dis- tinguish between benign and malignant characteris- tics of adrenal lesions, including computed tomography (CT) attenuation on an unenhanced scan and the pattern of enhancement, or washout, on enhanced scans. The risk of malignancy in- creases with tumor size; in tumors <4 cm, the re- ported risk of malignancy is approximately 2% and increases to >25% for lesions >6 cm.” Therefore, in the absence of other suspicious imag- ing characteristics, current American Association of
Clinical Endocrinologists (AACE)/American Asso- ciation of Endocrine Surgeons (AAES) guidelines for the management of adrenal incidentalomas rec- ommend the resection of adrenal lesions ≥4 cm in size. The National Comprehensive Cancer Network guidelines recommend adrenalectomy for incident- alomas >4-6 cm.15
ACC is a highly aggressive disease with a poor prognosis. At the time of diagnosis of ACC, 26.5% of patients will have nodal metastases, and 21.6% will have distant metastases. Reported 5-year sur- vival rates for patients undergoing resection is 38.6% (median survival, 31.9 months).16 The con- sequences of a missed ACC in small tumors may be significant; therefore, the true size threshold for resection of an adrenal incidentaloma is uncer- tain.1,3,4,6,10,17 The objective of this study was to as- certain the most cost-effective management algorithm in patients with nonfunctional, adrenal tumors, taking into consideration patient demo- graphics and clinical characteristics of the tumor.
METHODS
A Markov decision model was constructed to determine cost-effectiveness of surgical resection compared with either surveillance or no additional follow-up (“sign-off”) of a nonfunctional, inciden- tally identified adrenal mass (Fig 1). All relevant costs were considered from a societal perspective. Effectiveness was measured in units of life-years saved (LYS). An incremental cost-effectiveness
ratio (ICER) was calculated for each of the 3 alter- native management strategies. By convention, an ICER of <$50,000/LYS was considered to be the threshold for cost-effectiveness.18
The model was created using Tree-Age Pro (Williamstown, MA). Patients transitioned between mutually exclusive health states according to clin- ical probabilities. Each cycle in the model repre- sented 1 month; all costs and outcomes were adjusted to reflect this cycle length. Half-cycle corrections accounted for approximation errors. Patients cycled through the model until death (from natural causes, complications from surgery, or adrenal cortical cancer), or until the model reached a steady state.
Clinical probabilities and life expectancy of pa- tients with ACC were obtained from a review of the literature performed by 2 independent reviewers (TSW, KC; Tables I and II).19,20 Life expectancy of patients without ACC was culled from national life tables. Clinical probabilities for enlargement of ad- renal tumors and/or development of functionality were obtained from review of literature.21 Costs were derived from the 2010 Medicare Reimburse- ment schedule.22 Duration of hospital stay and asso- ciated costs were obtained from the Healthcare Cost and Utilization Project Nationwide Inpatient Sam- ple; as per convention, indirect costs were not in- cluded.23 All costs were reported in 2011 US dollars. A discount rate of 3% per year was applied to all future costs and outcomes.18
Nonfunctional 4cm adrenal incidentaloma
Sign-off
Surveillance
Surgery
No followup
Biochemical evaluation and/or CT scan
Benign
ACC
Patientre-presents with increase in size or function
No change
Change in size or functionality
Death from natural causes
Death from ACC
Surgery
ACC
Surgery
Benign
Palliation
Benign
ACC
Death from natural causes
Death from ACC
Death from natural causes
Death from ACC
| Input | Value | Reference |
|---|---|---|
| Costs ($) | ||
| Laparoscopic adrenalectomy* | 1,547.21 | 22 |
| Open adrenalectomy* | 1,753.85 | 22 |
| Biochemical evaluationt | 284.84 | 22 |
| Computed tomography# | 374.43 | 22 |
| Hospital stay (per night) | See Table II | |
| Annual discount rate | 3% | 18 |
| Clinical probabilities | ||
| Prevalence of ACC | 0.0096 | 3,4 |
| Likelihood ratio of ACC | 11 | |
| 2 cm | 1.01 | |
| 4 cm | 2.01 | |
| 6 cm | 4.4 | |
| 8 cm | 16.9 | |
| 10 cm | 24.4 | |
| Five-year survival with ACC | 20 | |
| Stage 1 | 0.75 | |
| Stage 2 | 0.64 | |
| Stage 3 | 0.58 | |
| Stage 4 | 0.15 | |
| Ten-year risk of hyperfunction | 21 | |
| Overall risk | 0.095 | |
| Tumor <3 cm | 0.033 | |
| Tumor ≥3 cm | 0.19 | |
| Ten-year risk of mass enlargement | 0.093 | 21 |
| Complication with laparoscopic adrenalectomy | 0.062 | 30 |
| Mortality with laparoscopic adrenalectomy | 0.003 | 30 |
| Complication with open adrenalectomy | 0.161 | 30 |
| Mortality with open adrenalectomy | 0.005 | 30 |
*Cost of surgery includes all costs associated with surgery, anesthesia, and pathology.
+Cost of biochemical evaluation includes serum aldosterone, renin, electrolytes, metanephrines, and cortisol (every year for 5 years). #Adrenal protocol computed tomography every 6 months for 2 years.
In the base-case model, a hypothetic cohort of patients with a 4-cm, nonfunctional, adrenal inci- dentaloma was randomized to undergo 1 of the 3 management strategies: surgery, surveillance, or sign-off. Patients with other known primary malig- nancies with potential adrenal metastases were excluded from the analysis. At the time of ran- domization, patients were assumed to be appro- priate candidates for adrenalectomy. In the surgery arm, all patients underwent laparoscopic resection, incurring the associated costs and risks of opera- tion. Patients with benign pathologic results con- tinued in the model with a normal life expectancy. Patients with ACC continued in the model until death, dependent on the stage of disease at the time of resection. A proportion of patients with ACC presented with advanced stage disease de- pendent upon clinical probabilities.19
In the surveillance arm, patients underwent the following surveillance tests: CT every 6 months for the first 2 years and annual biochemical testing for the first 5 years. If malignancy was suspected on
any of these serial imaging investigations, open adrenalectomy was performed. Life expectancy was determined by the stage of disease at the time of resection. For patients with ACC, delayed diagnosis resulted in progression of disease and a reduction in life expectancy. Patients in the sign-off arm did not undergo any additional follow-up in the model. If an ACC was missed, these patients later re-presented with advanced disease and were no longer considered candidates for surgery.
One-way sensitivity analyses were performed to test the impact of clinical uncertainty around model inputs in the base-case analysis. All clinical inputs were tested. Inputs were varied according to reasonable ranges found during literature review or, if these data were not available, by a +10% magnitude from the base case.
RESULTS
In the base-case analysis for a patient of 40 years, operation for an adrenal incidentaloma was more costly (4,822), but more effective
| 40 y | 41-60 y | 61-70 y | >70 y | |
|---|---|---|---|---|
| Laparoscopic | ||||
| Complication (%) | 6.2 | 7.9 | 8.5 | 21.9 |
| Mortality (%) | 0.3 | 0.1 | 0.5 | 2.9 |
| Hospital stay, no complication ($US) | 8,825 | 9,789 | 9,883 | 10,693 |
| Hospital stay, with complication ($US) | 14,538 | 16,833 | 15,362 | 17,553 |
| Open | ||||
| Complication (%) | 16.1 | 15.3 | 21.8 | 22.7 |
| Mortality (%) | 0.5 | 0.5 | 1.4 | 2.2 |
| Hospital stay, no complication ($US) | 10,755 | 11,532 | 11,902 | 12,816 |
| Hospital stay, with complication ($US) | 21,766 | 22,015 | 23,694 | 26,694 |
(477.33 vs 477.08 LYS) than surveillance, with an ICER of 2,958) but yielded significantly fewer life-years (244.69 LYS). Both surgery (8/LYS) were more cost-effective incrementally than sign-off.
Variations in clinical inputs resulted in changes in the cost-effectiveness analysis. Sensitivity analy- ses revealed that the model was susceptible to changes in the mortality rates of patients with ACC, the probability of ACC, the probability of compli- cations after laparoscopic surgery, the risk of advanced stage ACC, and the cost of surgery. and CT scan (Fig 2). However, the results of the base- case analysis were robust, and surgery remained more cost-effective than surveillance throughout the majority sensitivity analyses.
Improvements in survival resulted in operation becoming less cost-effective; varying mortality rates of patients with ACC by +10% from the base case resulted in ICERs ranging from 46,952/LYS (+10%). Lesser probabil- ities of ACC in a 4-cm tumor also resulted in decreased cost-effectiveness of the operation. ICER ranged from 30,185/LYS (-10%) for a +10% variation in probability of ACC. Surgery became less cost-effective than sur- veillance when the probability of ACC was <0.0125. Delayed diagnosis of ACC owing to
☒ ACC survival
☐ Probability of ACC
Cost/duration of hospital stay
Probability of complication with laparoscopic adrenalectomy
☐ Probability of mortality with laparoscopic adrenalectomy
☒ Probabilty of advanced stage ACC
☒ Cost of surgery
Cost of CT scan
☐ Cost of biochemical evaluation
$17K
$27K
$37K
$47K
Cost-effectiveness ($/LYS)
surveillance may result in patients presenting with a more advanced stage of disease. Surgery was more cost-effective with increased risks of advanced stage ACC, although sensitivity analyses revealed that the effect was small ($25,189- 26,532/LYS for a +10% variation from base case).
In the base-case, patients were assumed to pre- sent with a nonfunctional, 4-cm adrenal tumor and the probability of ACC and risk of advanced stage disease was directly dependent on adrenal tumor size based on the prevalence of ACC from the literature. Sensitivity analyses varying the size of the adrenal incidentaloma from 2 to 10 cm resulted in ICERs ranging from 84,947/ LYS (2 cm; Fig 3). For larger adrenal incidenta- lomas, laparoscopic adrenalectomy may not be ap- propriate owing to the increased risk of ACC. Sensitivity analyses utilizing the costs and attendant morbidity of open surgery for adrenal tumors ≥6 cm in size did not significantly affect the results.
The base-case analysis was a 40-year-old patient. Older patients undergoing surgery had increased rates of complications, duration of stay, and costs of hospitalization. Sensitivity analyses of patient age demonstrated that for patients >65 years old,
100,000
Cost-effectiveness ($/LYS)
80,000
60,000
Surgery not cost-effective
40,000
Surgery cost-effective
20,000
0
0
1
2
3
4
5
6
7
8
9
10
Incidentaloma size (cm)
100,000
Surveillance
Cost-effectiveness ($/LYS)
80,000
dominates
60,000
Surgery not cost-effective
40,000
Surgery cost-effective
20,000
0
0
20
40
60
80
Patient age (years)
surgery was not cost-effective compared with sur- veillance (ICER, $80,139/LYS; Fig 4). For patients >75 years of age, surveillance was less costly and more effective than surgery.
DISCUSSION
Although the majority of adrenal incidentalomas are nonfunctional, benign adenomas, given the aggressiveness of ACC, the consequences of a delay in diagnosis may have a significant effect on patient morbidity and mortality. Because the risk of malig- nancy increases with tumor size, current guidelines recommend resection of all adrenal tumors ≥4 cm in size. The results of this study show that this recommendation is a cost-effective approach and, in fact, adrenalectomy might be considered for patients with adrenal tumors ≥4 cm.
A review of the literature suggests that, for patients with incidental adrenal tumors, ≤25% may have mild hypercortisolism (“subclinical Cushing syndrome”), 0.5-3% may have hyperaldo- steronism, and 3-9% may have pheochromocyto- mas.1,9 Libe et al7 followed 64 patients with adrenal incidentalomas for a median of 25.5 months. Over this time period, 18 (28%) developed endocrine abnormalities, with a cumulative risk of 17% at 1 year, 29% at 2 years, and 47% at 5 years. The risk was greater during the initial 2 years for tu- mors ≥3 cm (23 vs 13% at 1 year; 39% vs 24% at 2 years; P = NS). However, in a prospective study of 229 patients, Bülow et al® found that, over a follow-up period of 27-30 months, only 3 (2%) pa- tients developed hormonal activity.8
Although the likelihood of a benign adrenal adenoma developing into ACC is likely rare, the true incidence is unknown, and it is known that adrenal tumors may grow over time. Therefore, current guidelines also recommend that adrenal
incidentalomas that do not meet initial criteria for resection undergo reevaluation radiographi- cally at 3-6 months and then annually for 1-2 years.4 Libe et al7 reported an increase in size of ≥1 cm (mean increment, 1.7) in 13 (20%) pa- tients; the cumulative risk of enlargement was 6% at 1 year, 14% at 2 years, and 29% at 5 years. The risk of growth was higher in patients with nor- mal adrenal function and within the first 2 years after identification of the adrenal incidentaloma.7 Bülow et al8 found that an increase of ≥0.5 cm was reported in 17 (7%) of patients and ≥1.0 cm in 12 (5%) of the 229 patients. Neither study reported the detection of ACC in patients who underwent adrenalectomy after an initial period of surveillance.7,8
Current recommendations for adrenalectomy based on tumor size are based on published literature, which has suggested that tumors ≥4 cm are at increased risk for malignancy.4,9 Size seems to be the single best indicator of malig- nancy; in an early report, 92% of ACC were found to be >6 cm in diameter.24 A multicenter, retro- spective analysis of 210 adrenal incidentalomas ini- tially reported that at a cutoff of 5 cm, there was a 93% sensitivity and 64% specificity in discriminat- ing between benign and malignant lesions.12 A re- cent study of 75 resected adrenal lesions with ≥2 imaging studies suggested that a change in size of 0.8 cm provided the greatest combined sensitiv- ity (60%) and specificity (84.6%) for differentiat- ing a benign from a malignant mass. Sturgeon et al11 published a risk assessment based on ACC recorded in the Surveillance, Epidemiology, and End Results database. The sensitivity, specificity, and likelihood ratios of tumor size to predict ma- lignancy were 91%, 80%, and 4.4, respectively, for tumors ≥6 cm, and 96%, 52%, and 2.0 for tumors ≥4 cm, suggesting that 4 cm is an appropriate threshold by which to recommend resection, based on risk of ACC alone.11
Although Sturgeon et al11 found that, for tu- mors ≥2 cm, the sensitivity, specificity, and likeli- hood ratio of tumor size predicting malignancy were 99%, 2%, and 1.0, respectively, ACC has been reported in tumors <4 cm, and professional society guidelines should be considered as a start- ing point for discussion of the indication for adrenalectomy, rather than be followed dogmati- cally.11,17,25 Complete surgical resection is the only potentially curative treatment for ACC, be- cause there are few effective adjuvant treatment modalities. Patients with a complete surgical resec- tion have reported 5-year survival rates of 40-50%; this decreases to a median survival of <1 year in
those patients who have an incomplete resection.26 Given the rarity and high mortality rates of ACC, there are few data regarding the natural history of untreated ACC; however, the potential conse- quences of a delay in diagnosis, leading to upstag- ing of ACC, could be significant. The results of this study suggest that size alone should not serve as the only criterion for operative resection; both age and cost of hospitalization were important var- iables in the cost-effectiveness of adrenalectomy.
Previous analyses of the costs of management of an adrenal incidentaloma have suggested that cur- rent recommendations for adrenalectomy are too aggressive, and that larger tumors, under certain circumstances, may be ignored. In a Dutch cost- effectiveness study, Kievit et al27 found that adrenal- ectomy was not cost-effective for any adrenal incidentaloma <6 cm. However, their pathways for diagnosis and treatment were based only on imag- ing and did not account for changes in hormonal ac- tivity or consequences of a missed ACC.27 A more recent European study suggested that given the rar- ity of ACC and long-term exposure to ionizing radi- ation from continued radiographic follow-up, current guidelines confer a risk of fatal cancer sim- ilar to the risk of adrenal malignancy. The study was not a true cost-analysis; furthermore, the au- thors only utilized a strategy of surgery or sign-off, which again precludes consideration of those pa- tients for whom adrenal tumors may either grow in size or become hyperfunctional.28
To our knowledge, this is the first cost-effectiveness study of the management of adrenal incidentalomas that employs current and widely utilized algorithms for both hormonal evaluation and radiographic imaging in the management of adrenal incidenta- lomas. Varying the probability of ACC on sensitivity analyses demonstrated that the results were robust; surgery remained cost-effective. Still, there are several limitations, including the lack of prospective data on the natural history of incidentalomas and of ACCand the need to make several assumptions based on the best quality of data available in the literature. The data utilized in the analyses for adrenal incidenta- lomas becoming hormonally active were based on available literature; however, there was not a consis- tent definition of “hormonal activity” or a standard- ized methodology for hormonal evaluation. Much of the current data is also based on surgical series, which may overestimate the prevalence of malignancy. Fur- thermore, for simplicity of the model, we chose not to include changes in radiographic appearance, such as border irregularity, necrosis, or contrast washout characteristics, as a criterion for surgery and assumed
that after initial detection, there was no change in tumor appearance, other than size, on imaging studies.
In summary, results of this analysis suggest that in patients with nonfunctional, incidentally de- tected adrenal tumors, surgical resection for le- sions >4 cm is a cost-effective approach. This is in accordance with current guidelines published by the AACE/AAES guidelines and more stringent than those of the National Comprehensive Cancer Network. Therefore, adrenalectomy should be considered in all tumors ≥4 cm after multidisci- plinary consideration of a patient’s medical history, surgical risk, and institutional and surgical outcomes.
The authors acknowledge Hadiza S. Kazaure, BSc, for her assistance in analysis of the Healthcare Cost and Utilization Project Nationwide Inpatient Sample data.29
REFERENCES
1. Zeiger MA, Siegelman SS, Hamrahian AH. Medical and sur- gical evaluation and treatment of adrenal incidentalomas. J Clin Endocrinol Metab 2011;96:2004-15.
2. Boland GW, Blake MA, Hahn PF, Mayo-Smith WW. Inciden- tal adrenal lesions: principles, techniques, and algorithms for imaging characterization. Radiology 2008;249:756-75.
3. NIH state-of-the-science statement on management of the clinically inapparent adrenal mass (“incidentaloma”). NIH Consensus State Sci Statements 2002;19:1-25.
4. Zeiger MA, Thompson GB, Duh QY, et al. The American As- sociation of Clinical Endocrinologists and American Associ- ation of Endocrine Surgeons medical guidelines for the management of adrenal incidentalomas. Endocr Pract 2009;15(Suppl 1):1-20.
5. Terzolo M, Stigliano A, Chiodini I, et al. AME position state- ment on adrenal incidentaloma. Eur J Endocrinol 2011; 164:851-70.
6. NCCN clinical practice guidelines in oncology: neuroendo- crine tumors. Version 1.2011 [cited 2012]. Available from: www.NCCN.org.
7. Libe R, Dall’Asta C, Barbetta L, Baccarelli A, Beck-Peccoz P, Ambrosi B. Long-term follow-up study of patients with adre- nal incidentalomas. Eur J Endocrinol 2002;147:489-94.
8. Bülow B, Jansson S, Juhlin C, et al. Adrenal incidentaloma: follow-up results from a Swedish prospective study. Eur J En- docrinol 2006;154:419-23.
9. Terzolo M, Bovio S, Pia A, Reimondo G, Angeli A. Manage- ment of adrenal incidentaloma. Best Pract Res Clin Endo- crinol Metab 2009;23:233-43.
10. Aron D, Terzolo M, Cawood TJ. Adrenal incidentalomas. Best Pract Res Clin Endocrinol Metab 2012;26:69-82.
11. Sturgeon C, Shen WT, Clark OH, Duh QY, Kebebew E. Risk assessment in 457 adrenal cortical carcinomas: how much does tumor size predict the likelihood of malignancy? J Am Coll Surg 2006;202:423-30.
12. Terzolo M, Ali A, Osella G, Mazza E. Prevalence of adrenal carcinoma among incidentally discovered adrenal masses. A retrospective study from 1989 to 1994. Gruppo Piemontese Incidentalomi Surrenalici. Arch Surg 1997;132:914-9.
13. Young WF, Jr. Management approaches to adrenal incident- alomas. A view from Rochester, Minnesota. Endocrinol Metab Clin North Am 2000;29:159-85.
14. Kasperlik-Zaluska AA, Otto M, Cichocki A, et al. Incidentally discovered adrenal tumors: a lesson from observation of 1,444 patients. Horm Metab Res 2008;40:338-41.
15. O’Neill CJ, Spence A, Logan B, et al. Adrenal incidenta- lomas: risk of adrenocortical carcinoma and clinical out- comes. J Surg Oncol 2010;102:450-3.
16. Bilimoria KY, Shen WT, Elaraj D, et al. Adrenocortical carci- noma in the United States: treatment utilization and prog- nostic factors. Cancer 2008;113:3130-6.
17. Dackiw AP, Lee JE, Gagel RF, Evans DB. Adrenal cortical car- cinoma. World J Surg 2001;25:914-26.
18. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA 1996;276:1253-8.
19. Paton BL, Novitsky YW, Zerey M, et al. Outcomes of adrenal cortical carcinoma in the United States. Surgery 2006;140: 914-20.
20. Lughezzani G, Sun M, Perrotte P, et al. The European Net- work for the Study of Adrenal Tumors staging system is prognostically superior to the international union against cancer-staging system: a North American validation. Eur J Cancer 2010;46:713-9.
21. Barzon L, Scaroni C, Sonino N, Fallo F, Paoletta A, Boscaro M. Risk factors and long-term follow-up of adrenal incident- alomas. J Clin Endocrinol Metab 1999;84:520-6.
22. Medicare reimbursement schedule [cited 2012 Feb 15]. Available from: http://www.cms.hhs.gov/home/rsds.asp.
23. Agency for Healthcare Research and Quality. HCUP NIS re- lated reports. NIS database documentation [cited 2012 Feb 15]. Available from: http://www.hcup-as.ahrq.gov/db/ nation/nis/nisdbdocumentation.jsp.
24. Copeland PM. The incidentally discovered adrenal mass. Ann Intern Med 1983;98:940-5.
25. Barnett CC, Jr, Varma DG, El-Naggar AK, et al. Limitations of size as a criterion in the evaluation of adrenal tumors. Surgery 2000;128:973-82.
26. Tacon LJ, Prichard RS, Soon PS, Robinson BG, Clifton-Bligh RJ, Sidhu SB. Current and emerging therapies for advanced adrenocortical carcinoma. Oncologist 2011;16:36-48.
27. Kievit J, Haak HR. Diagnosis and treatment of adrenal inci- dentaloma. A cost-effectiveness analysis. Endocrinol Metab Clin North Am 2000;29:69-90.
28. Cawood TJ, Hunt PJ, O’Shea D, Cole D, Soule S. Recommen- ded evaluation of adrenal incidentalomas is costly, has high false-positive rates and confers a risk of fatal cancer that is similar to the risk of the adrenal lesion becoming malignant; time for a rethink? Eur J Endocrinol 2009;161:513-27.
29. Kazaure HS, Roman SA, Sosa JA. Adrenalectomy in older Americans has increased morbidity and mortality: an analy- sis of 6,416 patients. Ann Surg Oncol 2011;18:2714-21.
30. Park HS, Roman SA, Sosa JA. Outcomes from 3144 adrenal- ectomies in the United States: which matters more, surgeon volume or specialty? Arch Surg 2009;144:1060-7.
DISCUSSION
Dr Quan-Yang Duh (San Francisco, CA): I like the fact that you used decision analysis to do this. I have a comment and a question.
The comment is this: The size criteria that we use of 4 and 6 cm. I think Demetrio Linos has looked into that, and the data come from pathologic examination of the
tumor that’s removed. And, in general, on CT scan, CT tends to underestimate the size. So many of us use 3 cm anyway, because the one that’s on CT, a 3-cm lesion, actually, when you take it out, looks like a 4-cm lesion.
My question for you is, did you take into account all potential complications? From what I can see, you assumed that the laparoscopic operation would go perfectly and only cost <$2,000. So did you take into account that if you have complications from the opera- tion, would that change your cost-effectiveness?
Dr Tracy Wang (Milwaukee, WI): Obviously, the cost of hospitalization would affect the cost-effectiveness. From the societal level, the numbers that we used from the nationwide inpatient sample are from a population of hospitals, which, you would assume, would include laparoscopic adrenalectomies that had longer hospital stays, or open adrenalectomies that had shorter hospital stays. So it’s a cohort of costs that we used.
Dr Michael J. Demeure (Scottsdale, AZ): At the risk of getting my colleagues from M.D. Anderson and Michi- gan up here, just to follow your logic, if the rationale for taking out these 4-cm tumors is that they could be malignant, why are you not doing open on all of them? And I think it’s safe from my end, at least, based on the German data and case-control studies, that if you can do a good operation, you can do it laparoscopically. So you’re sort of foraying into that logical minefield, I guess. I’d like you to comment on your perspective of why you do those laparoscopically.
Dr Tracy Wang: The model that we had was based on current recommendations, but, certainly, we fear that, over time, we may miss an adrenal cancer in a tumor. However, we were careful to exclude patients who, for any reason, had imaging characteristics that would auto- matically make you more worried for an adrenal cancer, such as irregular borders or necrosis.
In practice, certainly, if we had a 4-cm adrenal tumor, or a 3-cm adrenal tumor, that had any of those suspicious imaging characteristics, personally, I would tend to favor doing that open because of that higher risk. So this was really a study that was meant to say, well, if the current guidelines are recommending we remove 4-cm tumors because of the risk of malignancy, is that really a cost- effective approach?
If I understand your question correctly, our study was not exactly designed to answer that question, but more to examine what the current recommendations in the literature are.
Dr Clive Grant (Rochester, MN): Just as you were al- luding to was kind of the point of my question. And that is, you put great emphasis on the size criteria, 4 cm up or down a centimeter maybe; but what we talk about, and what you alluded to, is the imaging phenotype.
And I think that has a great deal, at least, to do in our decision making. If you have something that’s homoge- neous, hypodense, well circumscribed, and so on, you feel a good deal more secure with that being benign. And it also has to do with how things change.
So I think the comment that you had in your potential limitations is how those might change, but also the characteristics. I haven’t seen, obviously, your manuscript, but I think some of the things, like contrast uptake and some of the other characteristics, need to be really emphasized as part of that decision making.
Dr Tracy Wang: Certainly, we realize there are a vari- ety of different methods for performing abdominal CT scans when you’re looking for changes in adrenal wash- out and various other characteristics. And there is some literature, especially in the radiographic literature, that talks about changes over time that may influence the risk of adrenal cancer.
However, when we considered putting it in our model, it made what we were particularly looking at extremely complex. And so we chose to just completely eliminate that section of the analysis from this model, so that we could sort of simplify what we were trying to see as an initial study.
Dr Electron Kebebew (Bethesda, MD): One of the things I learned in San Francisco from Quan was to be more concerned about a patient who is 30 years old and has a 3- or 4-cm adrenal mass versus a 70-year-old. Were you able to include age and stratify the risk of ma- lignancy? That’s often, in many of the guidelines, not re- ally included, but an important variable.
Dr Tracy Wang: Although we performed a sensitivity analysis of age, we did not necessarily change the risk of malignancy based on patient age in our sensitivity analysis. So, for this, we assumed the same percentage of patients may have an adrenal cancer regardless of age.
Dr Cord Sturgeon (Chicago, IL): Can you tell us what you used to discount the future costs and qualities in the model, and also what the healthcare inflation rate was? We just saw a nice talk on how the costs of health care are rising.
Also, would your results have changed if the patient had to have an open adrenalectomy instead of a laparo- scopic? I perhaps missed that point; but, obviously, there’s a big difference in cost there. So you can tell us about that?
Dr Tracy Wang: We used an annual discount rate of 3%. A 3% annual discount rate is utilized in much of the cost- effectiveness literature. It does range between 3% and 5%. We have tended to stick with a 3% discount rate.
With respect to open versus laparoscopic adrenalec- tomy, again, to make some of our decision points a little bit simpler, in the model, we stuck to open adrenalec- tomy if there was suspicion of cancer. In the surveillance arm, for example, if a patient re-presented with a much larger tumor, they underwent open adrenalectomy.
Dr Cord Sturgeon: And healthcare cost inflation?
Dr Tracy Wang: We did not take that into account other than the discount rate that we used.
Dr Miguel Herrera (Mexico City, Mexico): Did you look at different cutoff sizes in your analysis? It seems to me that the cost analysis may not be different in differ- ent sizes. The reason being, for example, that small tu- mors are more likely to be found in young people, who may require a longer follow-up. Would you like to comment on that?
Dr Tracy Wang: We used the base case of 4 cm be- cause that is the recommendation for resection at this point in time. And, again, when we performed the sensi- tivity analysis based solely on size, our cost-effectiveness ranged; it was <$1,000 for a 10-cm tumor, and the incre- mental cost-effectiveness ratio was much higher for a 2- or 3-cm tumor.
So would a society would be willing to adopt, if you are willing to pay $100,000 to remove a 1-cm adrenal tumor and all associated risks, then than would be a cost- effective approach. So we did look at size in that respect.