Diagnosis and Management of Primary Hyperparathyroidism Across the Veterans Affairs Health Care System

Key Points

Question  How is primary hyperparathyroidism diagnosed and treated in a large integrated health system using electronic health records?

Findings  In this national cohort study of 371 370 veterans with chronic hypercalcemia, 86 887 (23.4%) were evaluated for primary hyperparathyroidism with serum parathyroid hormone level. Of 47 158 patients with hormonal evidence of primary hyperparathyroidism, 6049 (12.8%) underwent definitive treatment with parathyroidectomy.

Meaning  Primary hyperparathyroidism appears to be underdiagnosed despite the presence of chronic hypercalcemia, and it is undertreated with parathyroidectomy despite surgical indications.

Importance  Untreated primary hyperparathyroidism impairs quality of life and incurs substantial costs. Parathyroidectomy is a low-risk, high-success, definitive intervention.

Objectives  To determine the appropriateness of diagnostic evaluation for primary hyperparathyroidism in patients with hypercalcemia and the use of parathyroidectomy for the treatment of primary hyperparathyroidism across the Veterans Affairs (VA) health care system.

Design, Setting, and Participants  A retrospective cohort study of veterans with hypercalcemia and primary hyperparathyroidism was conducted from January 1, 2000, through September 30, 2015, using the VA Corporate Data Warehouse, a national electronic health record–based repository. The study included 371 370 veterans with chronic hypercalcemia and 47 158 veterans with biochemical evidence of primary hyperparathyroidism diagnosed by hypercalcemia, elevated serum parathyroid hormone levels, and near-normal serum creatinine levels. Statistical analysis was performed from April 21, 2017, to April 10, 2019.

Main Outcomes and Measures  The proportion of veterans with hypercalcemia who have parathyroid hormone levels evaluated, the proportion of veterans with hyperparathyroidism who are treated surgically, and the factors associated with parathyroidectomy using generalized linear latent and mixed model regression.

Results  Of 371 370 patients with chronic hypercalcemia, 86 887 (23.4%) received further testing with parathyroid hormone level. Of 47 158 patients meeting diagnostic criteria for primary hyperparathyroidism (42 737 men [90.6%] and 4421 women [9.4%]; mean [SD] age, 67.3 [11.8] years), 6048 (12.8%) underwent parathyroidectomy. Of 5793 patients with primary hyperparathyroidism presenting with a serum calcium level more than 1 mg/dL above the upper limit of normal, 1501 (25.9%) underwent parathyroidectomy. There was a decreasing trend in the use of parathyroidectomy over time. Factors positively associated with parathyroidectomy were nephrolithiasis (odds ratio [OR], 2.23; 95% CI, 1.90-2.61) and non-Hispanic white race/ethnicity (OR, 1.31; 95% CI, 1.17-1.46), while age (OR, 0.95; 95% CI, 0.95-0.96), Elixhauser Comorbidity Index score (OR, 0.76; 95% CI, 0.72-0.80), decreased estimated glomerular filtration rate (OR, 0.52; 95% CI, 0.45-0.60), and diagnosis of osteoporosis (OR, 0.65; 95% CI, 0.52-0.80) were inversely related to surgery.

Conclusions and Relevance  From this study’s findings, parathyroid hormone level is infrequently tested in patients with hypercalcemia, suggesting underdiagnosis of primary hyperparathyroidism. Patients meeting diagnostic criteria for primary hyperparathyroidism are undertreated with recommended parathyroidectomy. Similar gaps have previously been observed in non-VA care of primary hyperparathyroidism, suggesting the need for a systematic evaluation of barriers to diagnosis and treatment that informs intervention design.

Primary hyperparathyroidism (pHPT) is a common endocrine disorder in the United States, with an estimated incidence of 66 per 100 000 person-years among women and 25 per 100 000 among men.^1,2 Although pHPT is classically associated with nephrolithiasis, osteitis fibrosa cystica, and pathologic fractures, the clinical presentation of pHPT has evolved owing to newer diagnostic methods and routine laboratory testing allowing the diagnosis of early or asymptomatic disease.² Up to 80% of patients with pHPT now present before classic symptoms develop, and they have subtle manifestations of neuropsychiatric and cognitive disorders, musculoskeletal pain, gastrointestinal tract dysfunction, or osteoporosis.^2-4

Parathyroidectomy is the only definitive treatment for pHPT. It is a low-risk operation, commonly performed on an outpatient basis, with high cure rates and excellent long-term outcomes, including in elderly patients.^5-7 Surgical intervention for pHPT has been associated with increased bone mineral density, decreased risk of fracture, decreased progression of kidney disease, and decreased rates of associated cardiovascular dysfunction, constipation, and impotence.^8-14 Parathyroidectomy is the preferred treatment for all patients with symptomatic/classic disease, but some controversy remains about its use in patients with mild, asymptomatic pHPT.⁴ Accumulating evidence, though, shows benefits to patients undergoing parathyroidectomy for mild, asymptomatic disease, including improved neurocognitive and neuropsychiatric symptoms and overall quality of life.¹⁵ Cost benefits of parathyroidectomy, even for patients with asymptomatic disease, have also been demonstrated.^16-18 This favorable risk to benefit ratio of the procedure has made it the primary recommended treatment for pHPT, when surgery is possible.¹⁹ Management guidelines have been periodically updated and reflect the growing evidence on the benefits of parathyroidectomy.^20-23 The most recent management guidelines released by the American Association of Endocrine Surgeons suggest that parathyroidectomy should be considered in patients with mild, nonclassic symptoms, if surgical risk is not prohibitive.¹⁹

Single-institution and regional studies have shown that surgery is underutilized for pHPT across various practice settings.^24-27 The fragmented nature of health care in the United States may be partly accountable. Integrated systems with comprehensive electronic health records (EHRs), such as the Department of Veterans Affairs (VA), may be better equipped to identify at-risk patients, establish the correct diagnosis, and institute definitive treatment. Thus, we examined diagnostic evaluation for pHPT and treatment patterns in the VA health care system. Our study objectives were to assess the rate of diagnostic testing for pHPT among patients with hypercalcemia, examine the use of parathyroidectomy for pHPT and determine temporal trends, and identify factors associated with surgical treatment in pHPT.

We conducted a national retrospective cohort study of veterans with hypercalcemia between January 1, 2000, and September 30, 2015. The end date marks the official transition from International Classification of Diseases, Ninth Revision (ICD-9) coding to International Statistical Classification of Diseases and Related Health Problems, Tenth Revision coding at the VA. Data were obtained from the VA Corporate Data Warehouse,²⁸ an integrated electronic repository of patient data from both clinical and administrative systems across the United States. The study was approved by the Baylor College of Medicine Institutional Review Board, and a waiver of informed consent was granted because data were deidentified and presented as summary only.

We created a cohort of eligible patients by applying inclusion and exclusion criteria (Figure 1). After identifying adults (≥18 years of age) with hypercalcemia during the study period, we screened for patients with pHPT by applying a previously validated, laboratory-based algorithm: serum intact parathyroid hormone (PTH) level greater than the upper limit of normal and serum total calcium level greater than the upper limit of normal and serum creatinine level lower than 2.5 mg/dL (to convert to micromoles per liter, multiply by 76.25).²⁵ Values greater than the upper limit of normal were used in place of exact cutoffs to allow for variations in standard laboratory values between facilities. All 3 inclusion laboratory values were required to have resulted within a 6-month period. Patients with secondary (including renal) and tertiary hyperparathyroidism were excluded.

The rate of testing for pHPT among patients with hypercalcemia was assessed by calculating the proportion of patients with a serum PTH level among patients with 2 separate high serum calcium levels. Two separate levels were more likely to represent chronicity of the abnormality and exclude spurious laboratory test results. The overall rate of parathyroidectomy among patients with pHPT was determined based on Current Procedural Terminology codes for parathyroidectomy (60500, 60502, and 60505). An additional follow-up period of 6 months after the end of the study period was allowed when determining the rate of PTH testing, and any parathyroidectomy until December 1, 2018, was included. Patients who had undergone parathyroidectomy outside the VA under its coverage were identified within the VA Corporate Data Warehouse fee-basis domain, which captures outsourced medical care. A sensitivity analysis was performed to assess the rate of parathyroidectomy after excluding patients who had additional Medicare or Medicaid coverage or private insurance and could have received surgical care elsewhere.

To assess how guidelines for the management of pHPT may have influenced the use of surgery, we performed subgroup analyses assessing the rate of parathyroidectomy among patients having symptomatic/classic disease or meeting 1 or more surgical indications with asymptomatic disease. Symptomatic/classic disease included nephrolithiasis (ICD-9 codes, 592.0, 592.1, and 592.9) and/or pathologic fractures (ICD-9 codes, 733.1, 733.10-733.16, and 733.19). Surgical indications for asymptomatic disease included a serum calcium level greater than 1 mg/dL (to convert to millimoles per liter, multiply by 0.25) above the upper limit of normal, age younger than 50 years, estimated glomerular filtration rate less than 60 mL/min/1.73 m², 24-hour urine calcium level greater than 400 mg, and diagnosis of osteoporosis by ICD-9 codes (733.0, 733.00-733.03, and 733.09). For surgical patients, all ICD-9 codes and laboratory test values needed to be recorded prior to the date of surgery.

Covariates included demographic characteristics (age, sex, and race/ethnicity), laboratory data (serum calcium, creatinine, and PTH levels), and use of bisphosphonates and calcimimetics (cinacalcet). The presence of an ICD-9 code for hyperparathyroidism (252.0, 252.00-252.01, 252.08, 252.8, 252.9, and 227.1) 30 days or more prior to surgery was used as a surrogate marker of an established clinical diagnosis of pHPT. The burden of patients’ comorbid illnesses was quantified using the Elixhauser Comorbidity Index (eTable 1 in the Supplement).²⁹

Statistical analysis was performed from April 21, 2017, to April 10, 2019. The number of new diagnoses of pHPT based on laboratory test results was calculated for each year of the study period. The annual rate of parathyroidectomy was calculated as the number of operations performed each year divided by the number of new pHPT cases diagnosed that year. The nonparametric analysis of trend across ordered groups by Cuzick³⁰ was performed to determine if a significant trend in surgical intervention for pHPT existed during the study period.

Comparison of baseline characteristics, including demographic, laboratory, diagnostic, and administrative data, was performed with χ² or Fisher exact tests for categorical variables, t test for parametric continuous variables, and Wilcoxon rank sum test for nonparametric continuous variables. Hierarchical multilevel regression was used to identify factors associated with the use of parathyroidectomy while accounting for clustering of patients within hospitals. Variables with P < .20 on univariate analysis were included in the model (age, sex, race/ethnicity, number of comorbidities by Elixhauser Comorbidity Index, serum calcium level, serum PTH level, serum creatinine level, hyperparathyroidism ICD-9 code, estimated glomerular filtration rate <60 mL/min/1.73 m², nephrolithiasis, osteoporosis, and pathologic fracture). Because 24-hour urinary calcium level was available in only a small fraction of the cohort (n = 4953), we excluded it from the regression model. A sensitivity analysis was performed including the 24-hour urinary calcium level in the model to determine if inclusion had a significant association with the results.

All statistical analyses were conducted using Stata, version 14.0 (Stata Corp) with gllamm and nptrend extensions. All P values were from 2-sided tests and results were deemed statistically significant at P < .05.

A cohort of 47 158 patients with hormonal evidence of pHPT was created. The patients were predominantly male (42 737 [90.6%]) and white (29 170 [61.9%]), with a mean (SD) age of 67.3 (11.8) years. A total of 9324 patients with pHPT (19.8%) had an ICD-9 code for hyperparathyroidism in the EHR. Demographic characteristics and pHPT-relevant clinical, laboratory, and administrative data for the cohort are summarized in Table 1.

Of 866 349 patients with hypercalcemia, 371 370 had at least 2 separate high serum calcium values, and of those, 86 887 (23.4%) had a serum PTH level measured. Of all patients with hypercalcemia, 47 158 met inclusion criteria for pHPT. Of these patients, 6048 (12.8%) underwent parathyroidectomy. The number of new diagnoses of pHPT per year and the overall annual rate of parathyroidectomy for pHPT are shown in Figure 2. Although the number of new diagnoses each year increased from 790 in 2000 to 4673 in 2015, the proportion of patients receiving definitive treatment with parathyroidectomy declined from 21.6% in 2000 to 5.7% in 2015 (trend test; P < .001). The overall cure rate after parathyroidectomy was 93.7%, as reflected by a normal calcium level 6 months after surgery or later.

Subgroup analyses assessing the use of parathyroidectomy when surgical indications for asymptomatic or symptomatic/classic pHPT were specifically met are shown in Table 2. Low rates of parathyroidectomy were still observed for the subgroups of patients that met any of 2 surgical indications, which remained unchanged during the study period in management guidelines: 1501 of 5793 patients (25.9%) with a serum calcium level more than 1 mg/dL above normal and 774 of 3024 patients younger than 50 years (25.6%) underwent parathyroidectomy. Patients with elevated 24-hour urinary calcium levels (>400 mg) had the highest rate of operative intervention (128 of 189 [67.7%]), although 24-hour urinary calcium levels were infrequently measured (4953 of patients with pHPT [10.5%]). Low rates of parathyroidectomy were observed in patients with symptomatic/classic disease (426 of 1854 patients with nephrolithiasis [23.0%] and 12 of 204 patients with pathologic fractures [5.9%]).

Baseline health status and comorbidities did not appear to influence the decision for surgery. Even the healthiest patients (with no Elixhauser Index comorbidities), who additionally met 1 or more indications for surgery, underwent parathyroidectomy in only 17.7% of cases (2503 of 14 147 patients) (Table 2).

On sensitivity analysis, when patients with dual (VA and Medicare or Medicaid or private insurance) coverage were excluded, the overall use of parathyroidectomy among patients with pHPT was 12.6% (1540 of 12 207 patients).

On univariate analysis (Table 1), patients with pHPT who were treated surgically were significantly younger than the untreated patients (mean [SD], 60.8 [10.6] vs 68.2 [11.6] years; P < .001). Women (765 of 4421 [17.3%]), Hispanic patients (241 of 1501 [16.1%]), and non-Hispanic white patients (4012 of 29 170 [13.8%]) were more likely to be treated than men (5283 of 42 737 [12.4%]; P < .001) and other racial minorities (black, 1262 of 10 387 [12.1%]; and other, 533 of 6100 [8.7%]; P < .001). Patients treated with parathyroidectomy had fewer comorbidities. The parathyroidectomy cohort had higher serum calcium levels (median, 11.0 mg/dL; interquartile range [IQR], 10.7-11.4 mg/dL) and PTH levels (median, 120 pg/mL; IQR, 92-170 pg/mL [to convert to nanograms per liter, multiply by 1.0]) compared with the nonsurgical cohort (median calcium level, 10.7 mg/dL; IQR, 10.4-11.4 mg/dL; P < .001; and median PTH level, 100 pg/mL; IQR, 79-139 pg/mL; P < .001). A recorded clinical diagnosis of pHPT by ICD-9 code was significantly more common among patients undergoing parathyroidectomy (1577 of 6048 [26.1%] vs 7747 of 41 110 [18.8%]; P < .001). The use of bisphosphonates was higher in the surgical cohort (1339 of 6048 [22.1%]) compared with the nonsurgical cohort (5509 of 41 110 [13.4%]; P < .05). Among patients who underwent parathyroidectomy, 359 (5.9%) had received cinacalcet prior to surgery, whereas 2898 medically treated patients with pHPT (7.1%) received cinacalcet prior to surgery.

Using a generalized linear mixed model adjusting for the VA hospital where patients were treated, history of nephrolithiasis had the strongest association with operative intervention (odds ratio [OR], 2.23; 95% CI, 1.90-2.61) (Table 3). Other significant associations included non-Hispanic white race/ethnicity (OR, 1.31; 95% CI, 1.17-1.46), younger age (OR, 0.95; 95% CI, 0.95-0.96) and better preoperative health status as determined by number of comorbidities from the Elixhauser Comorbidity Index (OR, 0.76; 95% CI, 0.72-0.80). Diagnosis of osteoporosis (OR, 0.65; 95% CI, 0.52-0.80) and estimated glomerular filtration rate less than 60 mL/min/1.73 m² (OR, 0.52; 95% CI, 0.45-0.60) were associated with lower chance of parathyroidectomy.

On sensitivity analysis, elevated 24-hour urinary calcium level was also significantly associated with surgery (OR, 2.22; 95% CI, 1.40-3.51), when included in the regression model. The strength of association for all other variables remained unchanged except for non-Hispanic white race/ethnicity, which was no longer significant. This finding is likely secondary to a large proportion of patients for whom 24-hour urinary calcium values were available also identifying as non-Hispanic white.

Our national cohort study from the largest integrated health system in the United States reveals a low rate of diagnostic evaluation for pHPT among patients with hypercalcemia and underuse of curative surgical treatment, with only 12.8% of all patients with laboratory test findings diagnostic of pHPT undergoing parathyroidectomy. In addition, use of parathyroidectomy declined during the 15-year study period. Our study provides foundational data to guide improvements and underscores the need for practice changes to close these care gaps.

Testing for pHPT occurred in only 23.4% of patients with chronic hypercalcemia. Abnormal high levels of calcium are either missed or not followed up, or pHPT is not frequently considered in the differential diagnosis of hypercalcemia. Given the low rate of PTH testing in our cohort and previous reports suggesting that as many as 90% of patients with chronic hypercalcemia will have an underlying diagnosis of pHPT,¹ it is very likely that the actual prevalence of pHPT is even higher than observed in our population.

Our findings reveal substantial underuse of parathyroidectomy for the treatment of pHPT, even when proper testing is completed. This finding builds on prior single-institution and regional non-VA studies reporting the underuse of surgery in the treatment of pHPT, especially among elderly individuals.^24,26,31 Our findings use EHR data to further characterize the use of parathyroidectomy within a national, fully integrated health care system during a 15-year period. Underuse of parathyroidectomy remained significant, even when strong indications for surgery were present. Moreover, we observed a declining trend in the use of definitive surgical treatment. In the context of contemporary guidelines by the American Association of Endocrine Surgeons, which recommend consideration of parathyroidectomy for most patients with pHPT with either typical or atypical symptoms, these findings present an opportunity for system-wide improvement.¹⁹

Possible explanations for this care gap may include failure to notice the abnormal laboratory test results of hypercalcemia and high PTH level by the clinicians, inability to correctly interpret results and make a diagnosis of pHPT, or unfamiliarity with the guideline-recommended treatment of pHPT once the diagnosis is made. Only 19.8% of our pHPT cohort patients had an established diagnosis of hyperparathyroidism recorded in the EHR. A large academic health system also reported low rates of recorded pHPT diagnosis in the EHR despite presence of laboratory test result criteria for pHPT, with rates as low as 1.3%.²⁶

Lack of follow-up of abnormal test results has been linked to missed diagnoses across many disciplines, and comprehensive sociotechnical, EHR-based approaches have been suggested to address these issues.^32,33 The findings may also reflect a clinician knowledge gap related to existing evidence-based guidelines for the management of pHPT. Balentine et al³¹ showed that, although patients were properly evaluated and received a diagnosis of pHPT, only 29% had a referral for surgical evaluation. Moreover, Kuo et al²⁷ reported that only 23% of general practitioners, as opposed to 54% of endocrinologists, referred patients with pHPT for surgery, if 1 or more criteria were met. Only 31% of primary care clinicians in 1 non-VA survey reported that they were familiar with all the surgical indications for parathyroidectomy in patients with asymptomatic pHPT.³⁴ Primary care clinicians are usually the first to identify patients with hypercalcemia on routine laboratory testing, suggesting that additional detection strategies to identify patients with pHPT may be needed.

Access to high-quality surgical care may influence the decision for surgical referral. Experienced endocrine surgeons and otolaryngologists with a high volume of endocrine surgical procedures have better surgical outcomes.^35,36 Although we could not describe the availability of experienced endocrine surgery personnel in the VA, the surgical cure rates of 94% are comparable with the results of high-volume centers and suggest the presence of an adequate infrastructure to perform parathyroidectomies.³⁷

Patient-related factors that could not be captured in our study, such as refusal to undergo surgery, may also contribute to the low rate of surgery. Moreover, given that veterans have more chronic health conditions compared with the general population, it may be possible that patient comorbidities affected clinicians’ decision-making despite the safety and effectiveness of parathyroidectomy in older patients and in those at high risk for general anesthesia.^7,38,39 Nevertheless, the use of parathyroidectomy in our cohort was low even among patients with the lowest burden of comorbidities.

Our study calls for the development of novel strategies to identify missed opportunities in diagnosis and treatment. Developing accurate, EHR-based diagnostic algorithms (electronic triggers of a possible diagnosis generated by a constellation of EHR data) and implementing electronic clinical surveillance programs and novel clinical decision support tools to support clinicians have the potential to reduce missed opportunities in diagnosis and treatment and lead to improved quality of care.^40,41 Using these approaches, well-designed health informatics tools may help both primary care clinicians and specialists improve the management of pHPT at the VA and other health care systems.

Our findings are limited by the retrospective nature of our study and the use of EHR data without validation through medical record reviews. Second, patients with hypercalcemia with normal, but nonsuppressed, PTH levels, who studies have shown to also benefit from parathyroidectomy, were not considered in our study.⁴² If they had been included, the observed underdiagnosis of pHPT and underuse of surgery would have been higher. Similarly, we used total serum calcium level rather than corrected values of calcium because it is readily available. Third, our study spanned a prolonged period, during which surgical indications for asymptomatic disease underwent changes (eFigure 1 in the Supplement).^20-23 However, underuse of parathyroidectomy was observed even when surgical indications were constant (such as nephrolithiasis, fractures, serum calcium levels >1 mg/dL above normal, and age <50 years), suggesting that minor guideline changes did not affect the outcome. Fourth, although we have included data of outsourced care recorded by the VA, we may have missed parathyroidectomies performed under other insurance coverage, which may partly explain the decreasing trend in performing surgery for pHPT. Fifth, our findings may be limited by our predominantly male veteran population. However, the literature suggests that our findings are not VA specific, but rather reflect a wider problem of underdiagnosis and undertreatment of the disease.^25,27,31

On the other hand, the VA is the largest integrated health care system in the United States and appears to be the most suitable model to test our hypothesis at a national level. All patient records are accessible in a single EHR, even when patients relocate. Loss of patients to follow-up is generally low, referrals to specialists (eg, surgeons or endocrinologists) are unhindered, and health care costs to the individual patient are generally not prohibitive for surgery. Given these favorable characteristics of the VA and similar findings in smaller studies from different clinical practice settings, we believe that gaps in diagnosis and treatment of pHPT extend to other health care models, where the risks of fragmented care and miscommunication may be even greater.

Primary hyperparathyroidism appears to be underdiagnosed in a national veteran population and parathyroidectomy is underused for the treatment of the disease. Systematic evaluation of barriers to diagnosis and treatment is needed to design targeted interventions, including the use of health informatics tools and strategies to improve recognition and treatment.

Accepted for Publication: April 15, 2019.

Published Online: July 15, 2019. doi:10.1001/jamainternmed.2019.1747

Correction: This article was corrected on September 2, 2019, to fix an error in the title of Table 2 and the wording of the Conflict of Interest Disclosures.

Corresponding Author: Konstantinos I. Makris, MD, Michael E. DeBakey Department of Surgery, Baylor College of Medicine, 2002 Holcombe Blvd, OCL-112, Houston, TX 77030 (kmakris@bcm.edu).

Author Contributions: Drs Ramsey and Makris had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Alore, Suliburk, Massarweh, Balentine, Singh, Awad, Makris.

Acquisition, analysis, or interpretation of data: Alore, Ramsey, Massarweh, Balentine, Singh, Makris.

Drafting of the manuscript: Alore, Ramsey, Balentine, Singh, Makris.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Alore, Ramsey, Massarweh, Balentine, Awad.

Obtained funding: Makris.

Administrative, technical, or material support: Suliburk, Awad.

Supervision: Suliburk, Singh, Awad, Makris.

Conflict of Interest Disclosures: Dr Singh reported receiving grants from the Department of Veterans Affairs and grants from the Agency for Healthcare Research and Quality outside the submitted work. Dr Makris reported receiving grants from Baylor College of Medicine Department of Surgery and support from the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, and the Center for Innovations in Quality, Effectiveness and Safety (CIN 13-413), Michael E. DeBakey VA Medical Center, during the conduct of the study. No other disclosures were reported.

Funding/Support: The research reported here was supported in part by the US Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service, through the Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center. The study was also funded by the Seed Grant program at the Michael E. DeBakey Department of Surgery at Baylor College of Medicine.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the US Department of Veterans Affairs or the United States government.