Aldosterone-producing adenoma and other surgically correctable forms of primary aldosteronism
- 13k Downloads
Surgically correctable forms of primary aldosteronism are characterized by unilateral aldosterone hypersecretion and renin suppression, associated with varying degrees of hypertension and hypokalemia. Unilateral aldosterone hypersecretion is caused by an aldosterone-producing adenoma (also known as Conn's adenoma and aldosteronoma), primary unilateral adrenal hyperplasia and rare cases of aldosterone-producing adrenocortical carcinoma. In these forms, unilateral adrenalectomy can cure aldosterone excess and hypokalemia, but not necessarily hypertension. The prevalence of primary aldosteronism in the general population is not known. Its prevalence in referred hypertensive populations is estimated to be between 6 and 13%, of which 1.5 to 5% have an aldosterone-producing adenoma or primary unilateral adrenal hyperplasia. Taking into account referral biases, the prevalence of surgically correctable primary aldosteronism is probably less than 1.5% in the hypertensive population and less than 0.3% in the general adult population. Surgically correctable primary aldosteronism is sought in patients with hypokalemic, severe or resistant forms of hypertension. Recent recommendations suggest screening for primary aldosteronism using the aldosterone to renin ratio. Patients with a raised ratio then undergo confirmatory suppression tests. The differential diagnosis of hypokalemic hypertension with low renin includes mineralocorticoid excess, with the mineralocorticoid being cortisol or 11-deoxycorticosterone, apparent mineralocorticoid excess, pseudo-hypermineralocorticoidism in Liddle syndrome or exposure to glycyrrhizic acid. Once the diagnosis is confirmed, adrenal computed tomography is performed for all patients. If surgery is considered, taking into consideration the clinical context and the desire of the patient, adrenal vein sampling is performed to detect whether or not aldosterone hypersecretion is unilateral. Laparoscopic surgery for unilateral aldosterone hypersecretion is associated with a morbidity of about 8%, with most complications being minor. It generally results in the normalization of aldosterone secretion and kalemia, and in a large decrease in blood pressure, but normotension without treatment is only achieved in half of all cases. Normotension following adrenalectomy is more frequent in young patients with recent hypertension than in patients with long-standing hypertension or a family history of hypertension.
KeywordsAdenoma Aldosterone Primary Aldosteronism Eplerenone Glycyrrhizic Acid
This review deals with the prevalence, presentation, diagnosis and management of surgically correctable forms of primary aldosteronism (PA).
Disease name and synonyms
Primary aldosteronism subtypes
Surgically correctable subtypes:
Aldosterone-producing adenoma (alias Conn's adenoma, aldosteronoma), including:
Renin- or angiotensin-unresponsive adenoma
Renin- or angiotensin-responsive adenoma
Primary unilateral adrenal hyperplasia
Adrenocortical carcinoma with aldosterone hypersecretion
Non surgically correctable subtypes:
Idiopathic adrenal hyperplasia
Familial hyperaldosteronism type I (alias glucocorticoid-remediable aldosteronism), OMIM # 103900
Familial hyperaldosteronism type II, OMIM # 605635
Hyperaldosteronism is a condition caused by the overproduction of aldosterone, and is characterized by sodium retention and potassium excretion with resultant hypertension and hypokalemia. The condition was first described by J Conn , who further distinguished primary and secondary hyperaldosteronism on the basis of plasma renin levels, PA being characterized by renin suppression . In a recent clinical practice guideline, PA was defined as "a group of disorders in which aldosterone production is inappropriately high, relatively autonomous from the renin-angiotensin system, and non-suppressible by sodium loading" .
Prevalence of a raised aldosterone to renin ratio and of aldosterone-producing adenomas in referral samples or samples from a large community
Source and first author
ARR threshold, ng/dL per ng/mL.h-1
Subjects tested, n
Raised ARR, %
Confirmed PA, %
Proven adenoma, %
Community sample of hypertensive and non-hypertensive subjects
Newton-Cheh C, 2007
Referral samples of >1000 hypertensive patients
Nishikawa T, 2000
Rossi E, 2002
Rossi GP, 2006
Fogari R, 2007
Douma S, 2008
Adrenalectomy for Conn's adenoma has been reported in childhood . The prevalence of surgically correctable PA in children and adolescents is not known but is probably very low.
Between 1977 and 1981, the incidence rate of aldosterone-producing adenomas for which the patient underwent surgery was estimated at 0.8 per million individuals per year in Denmark . This figure is a low estimate, as it was obtained at a time when the aldosterone to renin ratio was not in use and computed tomography was not widely available. A national epidemiological survey in Japan estimated that 1,450 patients had been diagnosed with PA of any form in 1997 . In 1997, there were about 70 million Japanese adults aged 60 or less; thus, the incidence rate of PA could be estimated to be 2 cases per 100,000 individuals per year in this age group.
Patients with PA present with various degrees of hypertension and/or hypokalemia.
Surgically correctable PA is usually diagnosed in the fourth or fifth decade. Mean age at PA diagnosis across 9 large series ranged between 45 and 55 years, with an overall average of 50 years [15, 16, 17, 18, 19, 20, 21, 22, 23, 24]. Hypertension was generally detected 5 to 10 years before PA was diagnosed, indicating a significant delay between the onset of PA and its diagnosis. Grade III hypertension - with blood pressure (BP) levels of 180/110 mmHg or more - or resistant hypertension - with BP levels of 140/90 mmHg or more on triple antihypertensive treatment - are reported more frequently in patients with PA than in those with essential hypertension [7, 25]. The higher frequency of severe hypertension in patients with PA than in those with essential hypertension may be the consequence of an exploration bias, as current recommendations suggest screening for secondary causes, including PA, in patients with severe or difficult-to-treat hypertension . Left ventricular hypertrophy, microalbuminuria, and acute cardiovascular events are probably more frequent in patients with PA than in patients with essential hypertension and similar levels of BP (see Treatment objectives below).
Hypokalemia, usually defined as serum kalemia ≤ 3.5 mmol/l, is only present in a minority of patients with PA [5, 6, 7, 8, 9]. However, the frequency of hypokalemia is related to whether PA can be surgically cured. A study with a large series of patients reported that hypokalemia was present in 7%, 17% and 48% of patients with essential hypertension, idiopathic PA, and aldosterone-producing adenoma, respectively . Hypokalemia may be symptomatic and present as muscular weakness, cramps, paresthesia or palpitations with or without atrial fibrillation. There are rare cases in which PA is revealed by symptomatic hypokalemia without hypertension, with or without adenoma .
PA may also be documented in patients presenting with an incidentally detected adenoma. In a survey of 1096 patients with an adrenal 'incidentaloma', 16 patients were found to have PA, all of whom were moderately hypertensive .
Etiology of primary aldosteronism
The aldosterone to renin ratio is a heritable trait with a moderate degree of linkage to chromosome 11p . The etiology of aldosterone-producing adenoma and primary unilateral hyperplasia is not known.
Mechanism of hypertension in primary aldosteronism
The main effects of aldosterone are mediated by the mineralocorticoid receptor found in the cytosol of epithelial cells, particularly in the renal collecting duct. Aldosterone's major action on epithelial cells is to regulate the reabsorption of Na+, thereby also influencing the transport of water, K+, and H+ across the membrane. An electrochemical gradient permits the passage of sodium from the lumen into the epithelial cell through the amiloride-sensitive epithelial sodium channel. From there, active transport by the Na+/K+-ATPase carries the Na+ across the basolateral membrane, from the epithelial cell into the bloodstream, while simultaneously excreting K+; water follows the movement of the Na+. Aldosterone hypersecretion therefore increases exchangeable sodium, suppresses renin, increases the aldosterone to renin ratio, causes hypertension, and induces hypokalemia.
Diagnosis of primary aldosteronism
The aldosterone to renin ratio as a screening test
Aldosterone secretion increases when standing upright and with decreasing sodium intake. It decreases with decreasing plasma potassium concentrations and increasing age. Renin levels are also increased by decreases in sodium intake and the standing position, and they also decrease with increasing age. Therefore, using the aldosterone to renin ratio decreases the intra- and inter-patient variability in renin and aldosterone levels linked to sodium intake, body position and age[29, 30]. The aldosterone to renin ratio was introduced by K Hiramatsu and colleagues in 1981 as a screening tool to facilitate the diagnosis of PA among hypertensive patients .
Although logical and convenient, using the aldosterone to renin ratio has several limitations. First, renin levels can be determined as either plasma renin activity or active renin concentration, and aldosterone can be determined with iodinated or tritiated markers, with or without an extraction step. Consequently, reference values and diagnostic thresholds for renin, aldosterone and the aldosterone to renin ratio are laboratory-specific. Second, there is no agreement on the aldosterone to renin ratio cut-off value for diagnosing PA. In a systematic literature review, the aldosterone to plasma renin activity cut-off values suggested ranged from 7.2 to 100 ng/dl per ng/ml.h, corresponding to a 14-fold variation . The most frequently used cut-off values for aldosterone to plasma renin activity ratio are in the range of 20 to 50 ng/dl (554 to 1,385 pmol/l) per ng/ml.h; for aldosterone to active renin concentration, these values are in the range of 2.4 to 4.9 ng/dl (66 to 136 pmol/l) per mU/L (see  and Table 2). Third, the aldosterone to renin ratio is positively related to age, female sex, hypertensive status, and the use of beta-blockers or hormonal replacement therapy, and is negatively related to the use of angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers and diuretics. The aldosterone to renin ratio may be abnormally high in patients with normal aldosterone levels and very low renin levels, specifically in some elderly patients, some patients with a high sodium intake or those taking beta-blockers, in whom renin is undetectable. Some experts have therefore suggested that renin values below an appropriate minimal value (2.5 to 5 mU/l) should not be used to calculate the aldosterone to renin ratio [7, 22].
Some authors suggest screening for PA without the discontinuation of medication . However, antihypertensive agents alter the aldosterone to renin ratio. For example, in the Framingham Offspring Cohort, an aldosterone to renin ratio exceeding the value suggestive of PA was present in 3.1% of normotensive men and 8.8% of normotensive women, in 7.9% and 23.1% of untreated hypertensive men and women, and was present in 31.1% of men or women on beta-blockers (Figure 1). Therefore, diuretics and antihypertensive agents should be discontinued for at least two weeks and spironolactone, eplerenone and aliskiren for at least six weeks before determining the aldosterone to renin ratio. In cases in which a complete therapeutic washout would not be safe, antihypertensive medication should be limited to non-dihydropyridine calcium channel-blockers and alpha-blockers, which interfere minimally with the measurements. Hypokalemic patients are given potassium chloride for two reasons: to prevent arrhythmia and because hypokalemia inhibits aldosterone secretion, thereby increasing the risk of false negative results. In addition to antihypertensive agents, drospirenone, a progestin with antimineralcorticoid activity, may interfere with laboratory screening and confirmatory testing for the diagnosis of PAL and should be withdrawn in hypertensive women investigated for aldosteronism.
Renin-aldosterone dissociation is a key element of all definitions of PA. A raised aldosterone to renin ratio is a sensitive but non-specific test ([2, 3, 4, 5, 6, 7, 8, 9] and Table 2). Biochemical confirmatory tests are therefore necessary in patients with a positive aldosterone to renin ratio, to avoid costly and invasive imaging tests.
Suppression tests intended to confirm PA
Oral sodium loading test
Increase sodium intake to >200 mmol/d for 3 d, provide ClK to keep plasma K+
Urinary aldosterone determined from the morning of d 3 to the morning of d 4
PA unlikely if urinary aldosterone <10 μg/24 h
PA likely if urinary aldosterone >12 μg/24 h
Saline infusion test
Patient in recumbent position for at least 1 h, 2 liters of 0.9% saline iv over 4 h, starting at 0800-0930 h
Kalemia, aldosterone and cortisol at the beginning and the end of the test
PA unlikely if plasma aldosterone <5 ng/dl
PA likely if plasma aldosterone >10 ng/dl
Fludrocortisone suppression test
0.1 mg oral fludrocortisone every 6 h for 4 d. Provide slow-release KCl to keep plasma K+ and slow release NaCl to maintain urinary sodium excretion >3 mmol/kg body weight
Kalemia 4 times a day during the 4 days. On day 4
determine plasma cortisol, aldosterone and PRA in seated posture at 1000 h
PA likely if upright plasma aldosterone >6 ng/dl on day 4 at 1000 h
Captopril challenge test
25-50 mg captopril orally after sitting for at least 1 hour. Patient in seated position for 1 or 2 hours
Plasma aldosterone, PRA and cortisol before and 1 or 2 hours after captopril
PA likely if plasma aldosterone is not suppressed by captopril
Diagnosis of surgically correctable primary aldosteronism
As mentioned above, cases with unilateral aldosterone hypersecretion, including the classic aldosterone-producing Conn's adenoma, are surgically correctable forms of PA. Unilateral aldosterone hypersecretion should be confirmed by AVS in most or perhaps all cases (see  and Figure 2). Unfortunately, AVS is not widely available, and is an invasive test exposing patients to potential complications.
Non-secreting adenomas are present in about 2% of adult non-cancer patients and their prevalence increases with age . The presence of an adenoma in patients with PA suggests the presence of an aldosterone-producing adenoma, but it cannot exclude the combination of a non-secreting adenoma and idiopathic PA[3, 44]. The proportion of patients with a unilateral nodule on their computed tomography scan or magnetic resonance imaging, but a bilateral or contralateral secretion documented by AVS, was about 20% in a meta-analysis of 38 studies. Nonetheless, a non-secreting adenoma is unlikely in young patients with PA and the presence of an isolated characteristic adrenal adenoma > 1 cm in PA patients aged less than 40 [6, 50] or less than 55 is considered by some experts as an acceptable surrogate for diagnosing unilateral aldosterone hypersecretion.
Recent guidelines recommend computed tomography in all patients with confirmed PA to detect an adrenal carcinoma, even if an adrenalectomy is not otherwise considered. Adrenal carcinomas are very rare, with an annual incidence estimated at 1-2 per million population, and present as isolated PA in less than 5% of cases . Adrenal carcinomas presenting as PA are usually larger than 40 mm in diameter.
Adrenal vein sampling
AVS involves determining aldosterone and cortisol levels in the inferior vena cava and in the two adrenal veins. Considering the variability in ACTH secretion and the acute control of aldosterone secretion by ACTH, some experts advocate AVS during exogenous ACTH infusion. Other experts suggest AVS in the early morning, at the time of the spontaneous peak of ACTH secretion. A comparative study reported that exogenous ACTH infusion does not improve the detection of unilateral aldosterone hypersecretion if the two adrenal veins are catheterized simultaneously.
Proposed thresholds for interpreting results from adrenal vein sampling
Adrenal to IVC cortisol ratio
Dominant to non-dominant A/C ratio
Non-dominant to IVC A/C ratio
Adrenal venous sampling without ACTH stimulation
>4 or 5
Adrenal venous sampling with ACTH stimulation
AVS is a relatively complex procedure, with rates of failure (documented by a cortisol concentration in the cannulated vein(s) below two times higher than those in the inferior vena cava) between 3  and 22% . It is invasive, and carries a risk of complications of between 0.2  and 5% ; complications mainly include adrenal hematomas, groin hematomas and dissection of adrenal veins . However, AVS is superior to image-based techniques for therapeutic decisions, because the objective of surgery is to suppress unilateral hypersecretion, not a unilateral nodule. As mentioned above, a unilateral adenoma is compatible with an incidentaloma associated with idiopathic PA, particularly in elderly patients. Conversely, unilateral hypersecretion may be associated with primary unilateral adrenal hyperplasia undetectable on imaging. In a recent series, one in three patients undergoing adrenalectomy for PA had a lateralized aldosterone hypersecretion without a unilateral adenoma.
Hypertension with hypokalemia and suppressed renin is known as mineralocorticoid hypertension, the mineralocorticoid involved being aldosterone in the vast majority of cases . PA is easily excluded by the absence of aldosterone hypersecretion in cases of mineralocorticoid hypertension due to an excess secretion of cortisol or deoxycorticosterone, of which some can be corrected surgically; in apparent mineralocorticoid excess; during exposure to glycyrrhizic acid (liquorice); or in pseudo-hypermineralocorticoidism due to Liddle syndrome .
Management including treatment
Treatment objectives in patients with PA are to reduce BP, correct hypokalemia, and to prevent or reverse the eventual cardiovascular or renal alterations caused by aldosterone excess. In retrospective case-control studies, the cardiovascular and renal consequences of hypertension were reported to be more severe in patients with PA than in patients with essential hypertension and similar levels of office BP [63, 64]. Thus, correcting for aldosterone hypersecretion is a treatment objective per se . Whether PA is associated with an increased prevalence of glucose metabolism disorders is still disputed .
In patients with lateralized aldosterone hypersecretion, this goal can be achieved by adrenalectomy and probably by the long-term prescription of aldosterone antagonists. Patients' preferences should be taken into account. Candidates for surgery should be told that the presence of an aldosterone-producing adenoma poses no threat of cancer, that surgery may not cure their hypertension completely, and that the frequency of complications for laparoscopic adrenalectomy is about 8%.
With regard to the risk of hypokalemia-induced arrhythmia during anesthesia, hypokalemic patients should be provided potassium chloride or aldosterone antagonists before surgery.
A complete unilateral adrenalectomy is required in patients with primary unilateral adrenal hyperplasia. It is also preferable to adenoma enucleation in cases in which computed tomography has shown a Conn's adenoma, as multiple adenomas are frequent and are not necessarily identified by preoperative imaging. Besides, the adrenal gland is a small organ, and devascularizing an adenoma frequently results in the devascularization of the entire gland, making conservative surgery difficult. Laparoscopic surgery, using transperitoneal  or retroperitoneal approaches, is currently the procedure of choice. Mean operating time and length of hospitalization are typically 90 min and 4 days [22, 70], respectively. The mean complication rate is 8% [20, 70]. Complications of laparoscopic surgery include conversion to open surgery, hematoma due to intraoperative vascular injury, thromboembolism, pneumothorax or hemothorax, with most complications being benign.
The most common aspect is a unilateral, yellow, lipid-laden adenoma varying in diameter from 5 to 35 mm. Despite producing aldosterone, the tumor usually consists of zona fasciculata-type cells although zona glomerulosa- or mixed cell-type tumors have been described. Aldosterone-secreting adrenal carcinomas are extremely rare. These malignant tumors exceed 40 mm in size with involvement of local lymph nodes or invasion of adjacent organs .
Surgery abolishes aldosterone hypersecretion and hypokalemia in most patients with unilateral aldosterone hypersecretion [17, 20, 21, 22, 72]. It produces a large decrease in systolic BP (typically -20 to -40 mmHg), and in the number of antihypertensive medications prescribed (typically -1 therapeutic class) [19, 20, 21, 22].
Prediction of blood-pressure outcome
A dominant aldosterone-dependant component of hypertension and a low probability of associated essential hypertension are predictive of a better BP outcome after adrenalectomy.
In univariate analysis, several patient characteristics that suggest aldosterone-dependent hypertension have been associated with a favorable outcome: they include the presence of a typical and large aldosterone-producing adenoma on imaging studies or at pathological examination [15, 17, 20, 24, 72]; high urinary aldosterone excretion, low plasma renin or low serum potassium levels[18, 22, 72]; the absence of an increase in the plasma aldosterone concentration after standing up and the preoperative normalization of BP on monotherapy with high-dose spironolactone[17, 75]. Conversely, non-specific characteristics usually present in essential and/or severe hypertension have been associated with a poor BP outcome of adrenalectomy: they include older age [15, 17, 18, 19, 20, 72, 74, 75, 76] or longer duration of hypertension before surgery[17, 18, 19, 21, 22, 72]; higher body-mass index [21, 22, 72] male sex [15, 20, 72, 74]; a history of essential hypertension in first-degree relatives[17, 18]; preoperative BP[17, 21] and number of prescribed antihypertensive drugs [17, 18, 19, 20, 72]; and the presence of remodeling of resistance arteries.
Various factors have each been associated with a less favorable BP outcome in at least one multivariate model: lower urinary aldosterone excretion; a small adenoma or the presence of contralateral morphological abnormalities; the absence of BP control in patients on spironolactone; a higher number of antihypertensive medications required to control BP; older age or a longer history of hypertension; higher body mass index; male sex; and the presence of a family history of hypertension[15, 17, 18, 20, 22, 72]. The validity of these multivariate analyses is threatened by small sample sizes; comparisons across studies are difficult because the various models did not include the same variables.
The relevance of these prognostic markers for selecting patients for surgery should not be overemphasized, as they do not take into account the invariable cure of hypokalemia and hyperaldosteronism by adrenalectomy and provide only a weak prediction of the BP benefit in individual patients. For example, according to the only prediction model validated to date, number of antihypertensive medication ≤ 2, a body mass index ≤ 25 kg/m2, a duration of hypertension ≤ 6 years and female sex are the best predictors of hypertension cure following adrenalectomy. However, even if none of these features was present in an individual patient, this patient still had a 25% probability of being completely cured by an adrenalectomy and if not cured, hypertension was almost always better controlled. In another study in which the mean systolic BP decrease was -25 mmHg after surgery, the clinical impact of statistically significant prognostic factors was limited: the mean systolic BP decrease was only 3 mmHg less (-22 instead of -25 mmHg) in patients with a 0.5 mmol/l higher level of serum potassium before surgery, the most powerful predictor of unfavorable outcomes.
Operated patients can expect to be completely or partially weaned from mineralocorticoid antagonists or non-specific antihypertensive medication. The alternative to surgery is lifelong medication intended to correct or prevent the deleterious direct or indirect effects of hyperaldosteronism. Younger patients have a longer life expectancy and therefore derive a greater benefit from surgery. They also carry the smallest anesthetic risk. Early diagnosis of lateralized PA is therefore of paramount importance. The benefit-risk ratio is more balanced in older patients, especially if their antihypertensive medication has compelling indications, such as beta-blockers for coronary artery disease or angiotensin-converting enzyme and spironolactone for heart failure.
Alternatives to adrenalectomy
Mineralocorticoid receptor antagonists - spironolactone and eplerenone - provide a specific treatment for PA in patients who are not candidates for surgery. Unfortunately, only a few of these patients show a good BP response to spironolactone monotherapy . Furthermore, long-term tolerance of spironolactone at doses exceeding 50 mg per day is poor. There is no published evidence to suggest that high doses of eplerenone are more effective and better tolerated than spironolactone in patients with PA. If necessary, lower doses of aldosterone receptor antagonists may be associated with non-specific antihypertensive agents.
The etiopathogenesis and genomics of PA and aldosterone-producing adenomas are currently being studied by several research groups. Collaborative prospective studies are needed to document and standardize critical steps in the diagnosis of PA and the confirmation of lateralized PA in relation to surgically correctable PA. This specifically applies to studying the advantages that post-suppression plasma aldosterone concentrations have over basal concentrations, and the cut-off values that are used to detect a clinically relevant lateralizing ratio at AVS. As the rate of cure of hypertension following adrenalectomy is only 50%, there is also a need for randomized trials comparing the safety, acceptability and efficacy of surgery and aldosterone antagonists, regarding BP and target organ damage.
Surgically correctable PA is sought in patients with hypokalemic or difficult-to-treat hypertension, and is diagnosed by the presence of unilateral aldosterone hypersecretion at AVS. Surgery is particularly useful for young PA patients, who can be completely cured, and for PA patients with resistant hypertension, whose BP control can be markedly improved.
This study was supported in part by PHRC grant AOM 06 179 and by grants from INSERM and Ministère Délégué à la Recherche et des Nouvelles Technologies for the COMETE Network.
- 8.Fogari R, Preti P, Zoppi A, Rinaldi A, Fogari E, Mugellini A: Prevalence of primary aldosteronism among unselected hypertensive patients: a prospective study based on the use of an aldosterone/renin ratio above 25 as a screening test. Hypertens Res. 2007, 30: 111-7. 10.1291/hypres.30.111.CrossRefPubMedGoogle Scholar
- 43.Hirohara D, Nomura K, Okamoto T, Ujihara M, Takano K: Performance of the basal aldosterone to renin ratio and of the renin stimulation test by furosemide and upright posture in screening for aldosterone-producing adenoma in low renin hypertensives. J Clin Endocrinol Metab. 2001, 86: 4292-8. 10.1210/jc.86.9.4292.CrossRefPubMedGoogle Scholar
- 74.Stowasser M, Klemm SA, Tunny TJ, Storie WJ, Rutherford JC, Gordon RD: Response to unilateral adrenalectomy for aldosterone-producing adenoma: effect of potassium levels and angiotensin responsiveness. Clin Exp Pharmacol Physiol. 1994, 21: 319-22. 10.1111/j.1440-1681.1994.tb02520.x.CrossRefPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.