Although debate on the true prevalence of primary aldosteronism

Transcription

1 Annals of Internal Medicine Review Systematic Review: Diagnostic Procedures to Differentiate Unilateral From Bilateral Adrenal Abnormality in Primary Aldosteronism Marlies J.E. Kempers, MD, PhD; Jacques W.M. Lenders, MD, PhD; Lieke van Outheusden, MSc; Gert Jan van der Wilt, PhD; Leo J. Schultze Kool, MD, PhD; Ad R.M.M. Hermus, MD, PhD; and Jaap Deinum, MD, PhD Background: Computed tomography (CT), magnetic resonance imaging (MRI), and adrenal vein sampling (AVS) are used to distinguish unilateral from bilateral increased aldosterone secretion as a cause of primary aldosteronism. This distinction is crucial because unilateral primary aldosteronism can be treated surgically, whereas bilateral primary aldosteronism should be treated medically. Purpose: To determine the proportion of patients with primary aldosteronism whose CT or MRI results with regard to unilateral or bilateral adrenal abnormality agreed or did not agree with those of AVS. Data Sources: PubMed, MEDLINE, EMBASE, and Cochrane Library, 1977 to April Study Selection: Studies describing adults with primary aldosteronism who underwent CT/MRI and AVS were included. Of 472 initially identified studies, 38 met the selection criteria; extractable data were available for 950 patients. Data Extraction: The CT/MRI result was considered accurate when AVS showed unilaterally increased aldosterone secretion on the same side as the abnormality seen on CT/MRI or when AVS showed symmetric aldosterone secretion and CT/MRI revealed bilateral or no unilateral abnormality. Data Synthesis: In 37.8% of patients (359 of 950), CT/MRI results did not agree with AVS results. If only CT/MRI results had been used to determine lateralization of an adrenal abnormality, inappropriate adrenalectomy would have occurred in 14.6% of patients (where AVS showed a bilateral problem), inappropriate exclusion from adrenalectomy would have occurred in 19.1% (where AVS showed unilateral secretion), and adrenalectomy on the wrong side would have occurred in 3.9% (where AVS showed aldosterone secretion on the opposite side). Limitation: The lack of follow-up data in the included articles made it impossible to confirm that adrenalectomies were performed appropriately. Conclusion: When AVS is used as the criterion standard test for diagnosing laterality of aldosterone secretion in patients with primary aldosteronism, CT/MRI misdiagnosed the cause of primary aldosteronism in 37.8% of patients. Relying only on CT/MRI may lead to inappropriate treatment of patients with primary aldosteronism. Ann Intern Med. 2009;151: For author affiliations, see end of text. Although debate on the true prevalence of primary aldosteronism among the hypertensive population continues, primary aldosteronism is considered a frequent and curable form of hypertension. Depending on the population tested and the tools used to confirm the diagnosis, primary aldosteronism is reported to occur in approximately 5% to 13% of hypertensive patients, predominantly those with severe hypertension (1 8). Because patients with unilateral adrenal hypersecretion of aldosterone may be cured by unilateral adrenalectomy, differentiating unilateral (most often an adenoma) from bilateral (most often bilateral hyperplasia) aldosterone hypersecretion is important. In patients with bilateral hypersecretion, medical treatment, usually mineralocorticoid-receptor antagonists, is the therapy of choice. In the late 1960s, adrenal vein sampling (AVS) was introduced as a test to distinguish unilateral from bilateral primary aldosteronism (9). Later, computed tomography (CT) and magnetic resonance imaging (MRI) were adopted as the primary procedures with which to differentiate unilateral from bilateral adrenal abnormalities. In a considerable proportion of patients, however, CT and MRI results were found to disagree with those of AVS; CT/MRI would show, for example, a unilateral abnormal adrenal gland when aldosterone hypersecretion actually was occurring in the contralateral gland or bilaterally, or would show bilateral normal or abnormal adrenal glands when aldosterone hypersecretion was occurring in only 1 gland (10 14). A recent prospective study of 203 patients showed that operative planning based on anatomical imaging alone would have inappropriately excluded 21.7% of patients from adrenalectomy and would have led to unnecessary sugery in 24.7% (15). Thus, in recent years, AVS has regained popularity. Almost all experts agree that the criterion standard diagnostic test for lateralization of aldosterone secretion is measurement of aldosterone levels in the adrenal veins through AVS (11, 12, 16 20). However, AVS has not replaced CT/MRI because it is not universally available and CT/ MRI helps the surgeon to accurately localize the adrenal gland. In addition, AVS is a complicated technique with a See also: Print Editors Notes Editorial comment Web-Only CME quiz Conversion of graphics into slides 2009 American College of Physicians 329

2 Review Diagnostic Procedures to Differentiate Adrenal Abnormalities Context Primary aldosteronism can involve 1 or both adrenal glands. Surgery is indicated only for unilateral disease. Experts prefer adrenal vein sampling (AVS) to localize the source, but many physicians rely on computed tomography (CT) or magnetic resonance imaging (MRI). Content The authors reviewed 38 studies that compared localization by CT/MRI and AVS. In 37.8% of 950 patients, CT/ MRI results disagreed with AVS results. Based on CT/MRI alone, the following would have occurred: surgery for bilateral disease in 14.6% of patients, medical treatment for unilateral disease in 19.1%, and removal of the wrong adrenal in 3.9%. Caution Long-term outcomes, the best indicator of success, were often missing. Implication Because CT/MRI is not reliable, AVS is preferred for staging primary aldosteronism. The Editors relatively high chance of procedural failure (for example, when an adrenal vein cannot be adequately cannulated) in inexperienced hands and is more invasive than CT/MRI. Many physicians therefore perform CT/MRI as the first and sometimes only investigation to diagnose laterality of aldosterone secretion. Because misinterpretation of the imaging results could lead to inappropriate treatment, it is essential to know, under the assumption that AVS is the criterion standard test, how often inappropriate treatment decisions would have been made on the basis of CT/MRI findings alone. We performed a systematic literature search and analyzed studies of patients who underwent both techniques. METHODS Data Sources and Searches We conducted a systematic search of PubMed, MED- LINE, and EMBASE to find English-, French-, German-, or Dutch-language studies on primary aldosteronism by using the following search terms: ((primary hyperaldosteronism) OR(primary aldosteronism) OR(Conn) OR(hyperaldosteronism) OR (aldosterone-producing adenoma) OR (APA) OR(idiopathic hyperaldosteronism) OR(IHA) OR (primary adrenal hyperplasia) OR(PAH) OR(bilateral adrenal hyperplasia) OR(BAH)) AND ((adrenal venous sampling) OR(AVS) OR(venous sampling) OR(vein sampling) OR (adrenal vein) OR(adrenal venous)). We searched the Cochrane Library using the search string primary hyperaldosteronism or hyperaldosteronism and AVS or adrenal venous sampling for clinical trials published in English, French, German, or Dutch. Searches are up to date through April We sought to include all studies that performed both AVS and CT/MRI. We assumed that articles that reported on AVS results would also report on CT/MRI findings because CT/MRI is the standard imaging study in patients with primary aldosteronism. Study Selection Two reviewers independently and in duplicate assessed the eligibility of all abstracts. We excluded abstracts if they represented reviews or practice guidelines or if they, with certainty, described only studies in animals or children; patients without primary aldosteronism; or only 1 patient. If we could not make a decision about inclusion solely on the basis of the abstracts, we retrieved and reviewed fulltext articles. We considered studies eligible for inclusion if they met the following criteria: 1) original reported results, not previously published or used in earlier studies; 2) adult patients (age 18 years) with a diagnosis of primary aldosteronism; 3) description of more than 1 patient to avoid publication bias (when only 1 patient is described, there is a high a priori chance that the findings in this patient were unusual, which could imply that such studies are more likely to contain results in which CT/MRI did not agree with AVS results); 4) CT/MRI plus bilateral selective AVS performed in all patients, with the results of both investigations reported; and 5) publication from 1977 onward (CT has been available since that year). We excluded articles if 1) data had already been published (only the most recent publication was used), 2) inclusion bias was suspected because patients with concordant results seemed to have been selectively included or because selective examples of concordant and discordant CT/MRI and AVS results were presumed to have been given, and 3) only discordant results were described. To ensure interobserver consistency, differences in interpretation were resolved by consensus of the 2 reviewers, by group conferences with the other authors, or by referencing the original full-text article. Data Extraction and Quality Assessment We made a database of all studies eligible for inclusion. For each eligible study, we recorded aggregated results for the patients for whom the study reported the following: CT/MRI and AVS results; the techniques of CT, MRI, and AVS (such as slice thickness of the CT/MRI images); use of synthetic adrenocorticotropic hormone (ACTH) during AVS; and the AVS criteria used to determine whether aldosterone secretion was lateralized. For the 950 patients whose CT/MRI and AVS results could be retrieved, we recorded such characteristics as age, sex, blood pressure, and serum potassium level and biochemical variables (such as criteria for diagnosis of primary aldosteronism), as well as treatment strategy and patient follow-up if we could link those characteristics with certainty to these patients September 2009 Annals of Internal Medicine Volume 151 Number 5

3 Diagnostic Procedures to Differentiate Adrenal Abnormalities Review Because the diagnostic criteria for primary aldosteronism differ in their stringency, we subdivided the articles by whether the diagnosis of primary aldosteronism was based on a sodium chloride loading test (the most stringent criterion) or on the aldosterone renin ratio (less stringent criterion) or plasma aldosterone concentration plus plasma renin activity or plasma renin concentration (less stringent criterion). We also evaluated whether the report mentioned the cutoff value of both the selectivity criterion and the lateralization criterion. The selectivity criterion is used to determine whether blood was drawn selectively from the adrenal veins and not from an adjacent vein; this ratio is calculated for both the left and the right side and is expressed as the C AV /C IVC ratio: [cortisol] adrenal vein /[cortisol] inferior vena cava ; when this ratio exceeds a certain arbitrary cutoff value, sampling can be considered selective. The lateralization criterion is used to determine whether aldosterone hypersecretion was unilateral or bilateral by comparing aldosterone and cortisol concentrations in the 2 adrenal veins; this ratio, A/C ips /AC cont, is calculated as [aldosterone]/[cortisol] ipsilateral adrenal vein /[aldosterone]/[cortisol] contralateral adrenal vein (the ipsilateral adrenal vein is the one with the highest [aldosterone]/[cortisol] ratio). When this ratio exceeds a certain arbitrary cutoff value, aldosterone secretion is recorded as lateralized. When we were uncertain about how the authors confirmed the diagnosis of primary aldosteronism, we contacted one of the contributing authors to ask for details about the measurement of aldosterone and renin and whether a sodium chloride loading test had been performed. We also contacted the authors when the article did not specify the criteria used during AVS; of the 31 authors contacted, 22 (71%) responded and 13 (59% of the responders) could provide us with additional information. We excluded articles in which we could not confirm that the authors based the diagnosis of primary aldosteronism on a sodium chloride loading test, aldosterone renin ratio, or plasma aldosterone concentration plus plasma renin activity. Twenty-one articles reported on the success rate of AVS. Of the 976 procedures performed, 183 were unsuccessful (overall success rate, 81.3%). Data Synthesis and Analysis Assuming AVS is the diagnostic reference criterion test, we analyzed how many times the CT/MRI result agreed or disagreed with the AVS result. The CT/MRI result was considered accurate if 1) both CT/MRI and AVS showed unilateral adrenal abnormality on the same side or 2) AVS showed symmetrical aldosterone secretion (indicating bilateral adrenal abnormality) but CT/MRI showed bilateral or no unilateral abnormalities (some studies only specified that they found no unilateral abnormalities). The CT/MRI result was considered inaccurate if 1) AVS showed unilaterally increased aldosterone secretion and CT/MRI showed no or bilateral abnormalities; 2) AVS showed symmetrical aldosterone secretion and CT/MRI showed a unilateral abnormality; or 3) both CT/ MRI and AVS showed a unilateral abnormality, but on different sides. RESULTS Literature The literature search yielded 472 articles. After reading the abstracts, we excluded 251 articles; after reading the full texts, we excluded an additional 178 articles (Figure). After we had gathered additional information by contacting the authors, we included 38 of the 43 remaining arti- Figure. Study flow diagram. Articles from PubMed, MEDLINE, EMBASE, and Cochrane Library (n = 472) Abstracts screened Articles remaining (n = 221) Full text screened Articles remaining (n = 43) Articles remaining (n = 38) Studies describe a total of 2052 patients with primary aldosteronism Data on CT or MRI and AVS could be extracted for 950 patients Articles excluded (n = 251) Reviews or practice guidelines: 85 Studies did not include humans: 15 Studies did not include adults: 8 Studies not on primary aldosteronism: 86 Studies with only 1 patient: 57 Articles excluded (n = 178) Reviews or practice guidelines: 8 Not about primary aldosteronism: 4 Data could not be extracted for >1 patient: 12 No AVS or no CT or MRI performed or described: 66 No comparison of CT or MRI and AVS results in individual patients: 70 Previously published results: 4 Duplicates: 6 Inclusion bias: 3 Only discordant findings were described: 5 Articles excluded (n = 5) Not specified whether diagnosis of primary aldosteronism was based on sodium chloride loading test, aldosterone renin ratio, or plasma aldosterone concentration and plasma renin activity or concentration Details of excluded references are available on request. AVS adrenal vein sampling; CT computed tomography; MRI magnetic resonance imaging. 1 September 2009 Annals of Internal Medicine Volume 151 Number 5 331

4 Review Diagnostic Procedures to Differentiate Adrenal Abnormalities cles because either a sodium chloride loading test had been performed or biochemical confirmation of primary aldosteronism relied on the aldosterone renin ratio or plasma aldosterone concentration plus plasma renin activity or concentration. We excluded 5 reports that did not specify whether aldosterone renin ratio, plasma aldosterone concentration, plasma renin activity, or plasma renin concentration had been measured. Of the total group of 2052 patients described in these 38 articles, we could extract the CT/MRI and AVS results for 950 patients with primary aldosteronism. Characteristics of Included Patients In all included articles, the authors had either done a sodium chloride loading test or biochemically confirmed primary aldosteronism by using the aldosterone renin ratio or plasma aldosterone concentration plus plasma renin activity or concentration. In 21 of the 38 included studies, we could obtain the selected patients clinical and biochemical characteristics at the time of diagnosis of primary aldosteronism. Sex, age, mean blood pressure, and presence of hypokalemia were described for 332 (35% of total selected cohort), 422 (44%), 249 (26%), and 310 (32%) patients, respectively. Fifty-four percent of patients were male and 46% were female, the mean age was 50.8 years, the mean blood pressure was 157/97 mm Hg, and 52% of patients had hypokalemia (as defined by the authors). Tests Used to Diagnose Primary Aldosteronism and Criteria Used to Document AVS Table 1 shows the characteristics of the 38 articles (11 13, 15, 20 53), subdivided by the tests used to confirm the diagnosis of primary aldosteronism and by the description (if any) of the selectivity and lateralization criteria. In 13 articles (upper 2 blocks in Table 1), a sodium chloride loading test was performed to confirm the diagnosis of primary aldosteronism. The remaining studies (lower 2 blocks in Table 1) based the diagnosis of primary aldosteronism on the aldosterone renin ratio or on plasma renin concentration. Nine of the 13 studies that specified the use of a sodium loading test also reported the cutoff values for selectivity and lateralization. Most studies used a C AV /C IVC ratio to determine whether blood was obtained selectively from the adrenal veins during AVS. Cutoff values varied from greater than 1.1 to greater than 5. Ten studies reported the use of synthetic ACTH before or during the AVS procedure (this hormone is given to stimulate aldosterone secretion in the affected gland, to minimize stress-induced fluctuations in aldosterone secretion during AVS, and to maximize the gradient in cortisol from the adrenal vein to the inferior vena cava and thus confirm successful sampling of the adrenal veins). Of the 28 studies that mentioned the use of a lateralization criterion, 21 used an A/C ips /A/C cont ratio. Six of these studies used suppression of the A/C ratio of contralateral gland (A/C cont A/C IVC ) as an additional criterion for lateralization of aldosterone secretion. Proportion of CT/MRI Results Not Agreeing With AVS Results Table 2 summarizes the CT/MRI results for each category of AVS results. Table 3 presents these data for each included study individually. In 359 patients (37.8%), CT/MRI results were inaccurate (Table 3). In 139 patients (14.6%), AVS showed a bilateral and CT/MRI showed a unilateral abnormality; in 181 patients (19.1%), AVS showed a unilateral and CT/ MRI showed a bilateral or no abnormality; and in 37 patients (3.9%), AVS showed a unilateral abnormality on one side and CT/MRI showed it on the contralateral side. In the remaining 2 patients, AVS showed a unilateral abnormality, but the exact CT/MRI result was not specified (it was merely reported to be discordant with the AVS result). When AVS showed symmetrical aldosterone secretion, CT/MRI showed unilateral abnormalities on the left side in 43 cases (30.9%) and on the right side in 7 cases (5.0%); in 89 cases (64.0%), the side of the unilateral CT/MRI abnormality was not specified (Table 2). When AVS showed unilaterally increased aldosterone secretion, CT/ MRI showed no abnormalities in 77 patients (35.3%), bilateral abnormalities in 87 patients (39.9%), and contralateral abnormalities in 37 patients (17.0%); in 17 patients (7.8%), CT/MRI did not show a unilateral abnormality, but whether CT/MRI showed no or bilateral abnormalities was not specified. In the other 591 patients (62.2%), CT/MRI results were accurate in 355 (37.4%) for unilateral adrenal abnormality and in 236 (24.8%) for bilateral abnormality (Table 3). As shown in Table 2, when AVS revealed symmetric aldosterone secretion, CT/MRI showed no abnormalities (in 83 patients [35.2%]), bilateral abnormalities (in 100 patients [42.4%]), or no unilateral abnormality but no mention of whether CT/MRI showed no or bilateral abnormalities (in 53 patients [22.5%]). When a sodium chloride loading test was performed, the proportion of CT/MRI results that did not agree with AVS results was slightly but not significantly higher than when the diagnosis of primary aldosteronism was based on plasma aldosterone concentration plus plasma renin activity or concentration (39.5% vs. 36.0%; P 0.264, chisquare test) (Tables 1 and 3). The proportion of CT/MRI results not agreeing with AVS results was similar regardless of whether the study reported the selectivity and lateralization criteria for AVS (37.8% and 37.5%, respectively; P 0.94). DISCUSSION Because only patients with primary aldosteronism caused by unilateral adrenal abnormality benefit from adrenalectomy, the distinction between unilateral and bilateral adrenal abnormality is crucial for making decisions about treatment. However, several studies have reported that CT or MRI results disagree with AVS results in many September 2009 Annals of Internal Medicine Volume 151 Number 5

5 Diagnostic Procedures to Differentiate Adrenal Abnormalities Review Table 1. Biochemical and Adrenal Vein Sampling Criteria Used in Included Articles* Study, Year (Reference) Type of Aldosterone and Renin Measurement Sodium Chloride Test Selectivity Criterion Lateralization Criterion Nwariaku et al, 2006 (20) ARR (PAC and PRA) Oral/intravenous C AV /C IVC 3 A/C ips /A/C cont 4 Young et al, 2004 (15) ARR (PAC and PRA) Oral C AV /C IVC 5 A/C ips /A/C cont ; optimal cutoff is being investigated Stowasser and Gordon, 2004 (21) ARR (PAC and PRA) Fludrocortisone C AV /C IVC 3 A/C ips /A/C IVC 2 and A/C cont A/C IVC Bernini et al, 2002 (22) ARR (PAC and PRA) Intravenous C AV /C IVC 3 A/C ips /A/C cont 2, A/C cont A/C IVC Castro et al, 2002 (23) ARR (PAC and PRA) Intravenous C AV /C IVC 3; A/C AV /A/C IVC 2 A/C ips /A/C cont 3 Magill et al, 2001 (11) ARR (PAC and PRA) Oral C AV /C IVC 3 A/C ips /A/C cont 4 and A/C cont A/C IVC McAlister and Lewanczuk, 1998 (13) PAC and PRA Oral C AV C IVC A/C ips /A/C cont 5 definite, 3 suggestive Geisinger et al, 1983 (24) PAC and PRA Intravenous C AV /C IVC 3 Unilateral elevation of PAC or A/C Mulatero et al, 2008 (25) ARR (PAC and PRA) Intravenous C AV /C IVC 2 A/C ips /A/C cont 4, A/C ips /A/C cont 3 and A/C cont A/C IVC Harper et al, 1999 (12) PAC and PRA Intravenous C AV /C IVC 2 NR Hambling et al, 1993 (26) PAC and PRA Oral/fludrocortisone NR A/C ips A/C IVC and A/C cont A/C IVC Gleason et al, 1993 (27) PAC and PRA Intravenous NR NR Linde et al, 1979 (28) PAC and PRA Intravenous NR NR Toniato et al, 2006 (29) PAC and PRA No C AV /C IVC 1.1 A/C ips /A/C cont 2 Harvey et al, 2006 (30) ARR (PAC and renin) No Pre-ACTH C AV /C IVC 3; post-acth C AV /C IVC 5 A/C ips /A/C cont 4orPAC ips /PAC cont 10 Omura et al, 2006 (31) PAC and PRA No Pre-ACTH C AV 1103 nmol/l; post-acth C AV 5518 nmol/l Pre-ACTH PAC 6.93 nmol/l; post-acth PAC nmol/l Carr et al, 2004 (32) PAC and PRA No C AV /C IVC 2 A/C ips /A/C cont 4 5 Castro et al, 2002 (23) ARR (PAC and PRA) No C AV /C IVC 3; A/C AV /A/C IVC 2 A/C ips /A/C cont 3 Glodny et al, 2000 (33) ARR (PAC and renin) No C AV /C IVC 3 A/C ips /A/C cont 2.5 Sheaves et al, 1996 (34) PAC and PRA No C AV C IVC A/C ips /A/C cont 5 Blumenfeld et al, 1994 (35) ARR (PAC and PRA) No C AV /C IVC 2 A/C ips A/C IVC and A/C cont A/C IVC White et al, 2008 (36) ARR (PAC and PRA) No C AV /C IVC 3 A/C ips /A/C cont 4 Schwab et al, 2008 (37) ARR No C AV /C IVC 3 A/C ips A/C cont and A/C cont A/C IVC Minami et al, 2008 (38) ARR (PAC and PRA) No C AV 5518 nmol/l A/C ips /A/C cont 4, /C cont A/C IVC ; PAC nmol/l on 1 side Satoh et al, 2007 (39) PAC and PRA No C AV /C IVC 5 A/C ips /A/C cont 3 Wu et al, 2001 (40) ARR (PAC and PRA) No NR PAC ips /PAC cont (after MCP) 5 definite, 3 suggestive Wu et al, 2008 (41) ARR (PAC and PRA) No NR A/C ips /A/C cont 2 and A/C cont /A/C IVC 1 Wang et al, 2000 (42) ARR (PAC and PRA) No NR NR Lo et al, 1996 (43) ARR (PAC and PRA) No NR NR Rao et al, 1995 (44) PAC and PRA No NR NR Naruse et al, 1994 (45) PAC and PRA No NR PAC ips /PAC cont 3 and/or A/C ips /A/C cont 3 Dunnick et al, 1993 (46) PAC and renin No NR A/C ips /A/C cont 3 Takasaki et al, 1987 (47) PAC and PRA No NR NR Ma et al, 1986 (48) PAC and PRA No NR A/C ips A/C IVC and A/C cont A/C IVC Witzgall et al, 1983 (49) PAC and PRA No C AV /C IVC 2 NR Dunnick et al, 1982 (50) PAC and renin No NR A/C ips /A/C cont 3 McAreavey et al, 1981 (51) PAC and PRA No No rigid cutoff value Optimal ratio and cutoff value are being investigated Nocaudie-Calzada et al, ARR (PAC and PRA) No NR A/C ips /A/C cont (52) Volpe et al, 2008 (53) PAC and PRA No C AV /C IVC 3 NR A aldosterone; ACTH adrenocorticotropic hormone; ARR aldosterone renin ratio; AV adrenal vein; C cortisol; cont contralateral (nondominant); ips ipsilateral (dominant); IVC inferior vena cava; MCP metoclopramide; NR not reported; PAC plasma aldosterone concentration; PRA plasma renin activity. * The table is divided into blocks; each block is indicated by a heavy black line. In the upper 2 blocks (Nwariaku et al through Linde et al), a sodium chloride loading test was performed to confirm the diagnosis of primary aldosteronism. In the lower 2 blocks (Toniato et al through Volpe et al), the diagnosis of primary aldosteronism was based on the ratio of plasma aldosterone concentration and plasma renin activity or plasma renin concentration. In the first and third blocks, the use of cutoff values for selectivity and lateralization was specified. With use of synthetic adrenocorticotropic hormone. One patient with sodium chloride loading test and 1 patient without sodium chloride loading test. Author indicated that sodium chloride loading tests were done in some patients, but data could not be traced to individuals. Two of the 3 criteria had to be met. 1 September 2009 Annals of Internal Medicine Volume 151 Number 5 333

6 Review Diagnostic Procedures to Differentiate Adrenal Abnormalities Table 2. Overview of CT/MRI Results When AVS Showed Unilateral or Bilateral Abnormality* CT/MRI Result AVS Result Accurate CT or MRI, % (n/n) Unilateral Right Unilateral Left Unilateral, Side Undefined Bilateral Unilateral 66.9 (355/531) Right Left Side undefined 89 Ipsilateral 239 Contralateral 29 Not unilateral (53/70) Bilateral 52.7 (183/347) Normal Abnormal Discordant unspecified 2 0 (0/2) Abnormalities accurately detected by CT or MRI, % (n/n) 68.1 (60/88) 69.1 (56/81) 58.9 (239/406) 62.9 (236/375) Overall: 62.2 (591/950) AVS adrenal vein sampling; CT computed tomography; MRI magnetic resonance imaging. * Unless otherwise noted, data are the number of patients. Inaccurate results by CT or MRI. patients (9 14). To learn more about the proportion of patients in whom a potentially wrong decision could be made by relying solely on the CT/MRI result, we performed a systematic literature review and analyzed studies that performed both techniques on the same patients. We found a strikingly high rate (37.8%) of CT/MRI results that did not agree with AVS results. Assuming that AVS is 100% accurate in detecting lateralization of aldosterone secretion, the hypothetical sole reliance on CT/ MRI to determine treatment strategy would have resulted in inappropriate adrenalectomy in 14.6% of patients, inappropriate exclusion from adrenalectomy in 19.1%, and adrenalectomy on the inappropriate side in 3.9%. Both CT/MRI and AVS showed a unilateral abnormality in slightly more than half of the patients (56.0% and 60.5%, respectively) but agreed on the involved side in the same patient in only 37.4%. Some patients did not undergo a sodium chloride loading test, which is the diagnostic reference standard test for primary aldosteronism, and some of them might not have primary aldosteronism. In theory, inadvertently including patients without primary aldosteronism (who will not show lateralization on AVS or a unilateral adrenal mass on CT/MRI) should increase the proportion of participants in whom CT/MRI and AVS results agree. In fact, the proportion of CT/MRI results that agreed with AVS results was the same in patients who underwent a sodium chloride test and patients who did not. Discrepancies between AVS and CT/MRI results are usually attributed to false-positive or false-negative CT/MRI results because CT/MRI might detect adenomas or nodules that are nonfunctioning and because CT/MRI can miss small adenomas. A precise comparison between AVS and CT/MRI is highly dependent on how carefully the CT/MRI scan has been read. Conceptually, AVS is an attractive technique because it detects the origin of abnormality on a biochemical basis in a primarily biochemical disease, such as primary aldosteronism. Indeed, few authors have reported false-positive AVS findings of unilateral abnormality (12, 13, 49, 54). However, studies might underestimate the occurrence of false-negative AVS results because patients do not undergo adrenal surgery for supposed bilateral hyperplasia. It is therefore not possible to confirm that the decision to refrain from adrenalectomy has been appropriate. Surely, the decision not to operate on these patients is appropriate because AVS is considered to be the criterion standard test, but studies claiming nearly 100% accuracy of AVS or CT/MRI for unilateral primary aldosteronism warrant a critical reappraisal (10, 24, 55). The large number of excluded studies reflects the wide variation in diagnostic approaches, the lack of diagnostic rigor, and the inconsistencies in CT/MRI reporting. In this context, 2 issues are important. First, the interpretation of AVS results might depend on whether ACTH was used during the AVS procedure because ACTH stimulation of aldosterone secretion in the affected gland is thought to result in better detection of unilateral abnormality. Second, the cutoff value used in the lateralization criterion will influence how accurately AVS can detect unilateral or bilateral abnormality (56). Our review showed a large variation in the lateralization criteria and the cutoff value used to diagnose unilateral secretion. We therefore could not evaluate the accuracy of a higher A/C ips /A/C cont ratio in diagnosing unilaterally increased aldosterone secretion. The recently issued guideline (16) on diagnosis and treatment of primary aldosteronism may improve the diagnosis of primary aldosteronism. When we designed this study, we hoped to further support conclusions about the accuracy of CT/MRI by obtaining pathologic confirmation of the diagnosis and follow-up data on patients who underwent adrenalectomy for adenoma. Although pathologic data were available for September 2009 Annals of Internal Medicine Volume 151 Number 5

7 Diagnostic Procedures to Differentiate Adrenal Abnormalities Review many patients who had surgery, information on biochemical measures, blood pressure, and medication use during follow-up was too sparse to justify including them in our report. We therefore did not further analyze these data. In our opinion, they should be recorded carefully in future studies on primary aldosteronism because they are indispensable for determining whether the decision to perform adrenalectomy was appropriate. On the basis of the poor performance of CT/MRI when AVS is used as the criterion standard test, we ques- Table 3. Proportion of CT/MRI Results Agreeing or Not Agreeing With Result of AVS* Study, Year (Reference) CT or MRI Agreeing With AVS CT or MRI Not Agreeing With AVS Agree/Not Agree, n/n AVS Bilateral and CT or MRI Bilateral or Normal, n AVS Unilateral and CT or MRI Unilateral, n Agree/Not Agree, n/n AVS Bilateral and CT or MRI Unilateral, n AVS Unilateral and CT or MRI Bilateral or Normal, n Nwariaku et al, 2006 (20) 22/ / Young et al, 2004 (15) 100/ / Stowasser and Gordon, 43/ / (21) Bernini et al, 2002 (22) 3/ / Castro et al, 2002 (23) 1/ / Magill et al, 2001 (11) 21/ / McAlister and Lewanczuk, 5/ / (13) Geisinger et al, 1983 (24) 20/ / Mulatero et al, 2008 (25) 55/ / AVS Ipsilateral and CT or MRI Contralateral, n Harper et al, 1999 (12) 4/ / Hambling et al, 1993 (26) 8/ / Gleason et al, 1993 (27) 8/ / Linde et al, 1979 (28) 4/ / Toniato et al, 2006 (29) 4/ / Harvey et al, 2006 (30) 19/ / Omura et al, 2006 (31) 55/ / Carr et al, 2004 (32) 12/ / Castro et al, 2002 (23) 0/ / Glodny et al, 2000 (33) 3/ / Sheaves et al, 1996 (34) 13/ / Blumenfeld et al, 1994 (35) 9/ / White et al, 2008 (36) 27/ / Schwab et al, 2008 (37) 8/ / Minami et al, 2008 (38) 20/ / Satoh et al, 2007 (39) 56/ / Wu et al, 2001 (40) 12/ / Wu et al, 2008 (41) 9/ / Wang et al, 2000 (42) 9/ / Lo et al, 1996 (43) 9/ / Rao et al, 1995 (44) 2/ / Naruse et al, 1994 (45) 1/ / Dunnick et al, 1993 (46) 12/ / Takasaki et al, 1987 (47) 1/ / Ma et al, 1986 (48) 2/ / Witzgall et al, 1983 (49) 3/ / Dunnick et al, 1982 (50) 4/ / McAreavey et al, 1981 (51) 1/ / Nocaudie-Calzada et al, 1/ / (52) Volpe et al, 2008 (53) 5/ / Total, n/n (%) 591/950 (62.2) 236/950 (24.8) 355/950 (37.4) 359/950 (37.8) 139/950 (14.6) 181/950 (19.1) 37/950 (3.9) AVS adrenal vein sampling; CT computed tomography; MRI magnetic resonance imaging. * The table is divided into blocks; each block is indicated by a heavy black line. In the upper 2 blocks (Nwariaku et al through Linde et al), a sodium chloride loading test was performed to confirm the diagnosis of primary aldosteronism. In the lower 2 blocks (Toniato et al through Volpe et al), the diagnosis of primary aldosteronism was based on the ratio of plasma aldosterone concentration and plasma renin activity or plasma renin concentration. In the first and third blocks, the use of cutoff values for selectivity and lateralization was specified. One patient with and 1 patient without a sodium chloride loading test, reported in the same study. We assumed that the data concern a different patient set, but this is not certain. In 2 patients in whom AVS showed lateralization on 1 side, CT and MRI were reported to give discordant results. However, the study did not specify whether the abnormality lateralized contralaterally or showed bilateral or no abnormalities. 1 September 2009 Annals of Internal Medicine Volume 151 Number 5 335

8 Review Diagnostic Procedures to Differentiate Adrenal Abnormalities tion the value of CT/MRI as an acceptable alternative to AVS. This conclusion may have major consequences for facility planning in centers that still rely on CT/MRI, especially in view of the high prevalence of primary aldosteronism. However, AVS is technically unsuccessful in about 20% of cases, mostly because of failure to cannulate the right adrenal vein. In these patients, CT/MRI is the only alternative, although it is imperfect, for diagnosing adenoma. When care must proceed on the basis of CT/MRI results alone, the physician should remember that CT/ MRI is very inaccurate. In conclusion, in 37.8% of patients with primary aldosteronism, the results of CT/MRI disagreed with the AVS results. If AVS is the criterion standard test of laterality of aldosterone secretion, proceeding on the basis of CT/MRI results alone should lead to an incorrect treatment decision to operate or not (or at which side) in 37.8% of cases. From Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands. Acknowledgment: The authors thank the authors of previous articles who provided us with additional information. Potential Financial Conflicts of Interest: None disclosed. Requests for Single Reprints: Jaap Deinum, MD, PhD, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, the Netherlands; , Current author addresses are available at References 1. Schwartz GL, Turner ST. Screening for primary aldosteronism in essential hypertension: diagnostic accuracy of the ratio of plasma aldosterone concentration to plasma renin activity. Clin Chem. 2005;51: [PMID: ] 2. Rossi GP, Bernini G, Caliumi C, Desideri G, Fabris B, Ferri C et al. A prospective study of the prevalence of primary aldosteronism in 1,125 hypertensive patients. J Am Coll Cardiol. 2006; 48: PAPY Study Investigators. A prospective study of the prevalence of primary aldosteronism in 1,125 hypertensive patients. J Am Coll Cardiol. 2006;48: [PMID: ] 3. Mosso L, Carvajal C, González A, Barraza A, Avila F, Montero J, et al. Primary aldosteronism and hypertensive disease. Hypertension. 2003;42: [PMID: ] 4. Young WF Jr. Minireview: primary aldosteronism changing concepts in diagnosis and treatment. Endocrinology. 2003;144: [PMID: ] 5. Fardella CE, Mosso L, Gómez-Sánchez C, Cortés P, Soto J, Gómez L, et al. Primary hyperaldosteronism in essential hypertensives: prevalence, biochemical profile, and molecular biology. J Clin Endocrinol Metab. 2000;85: [PMID: ] 6. Gordon RD, Ziesak MD, Tunny TJ, Stowasser M, Klemm SA. Evidence that primary aldosteronism may not be uncommon: 12% incidence among antihypertensive drug trial volunteers. Clin Exp Pharmacol Physiol. 1993;20: [PMID: ] 7. Mulatero P, Stowasser M, Loh KC, Fardella CE, Gordon RD, Mosso L, et al. Increased diagnosis of primary aldosteronism, including surgically correctable forms, in centers from five continents. J Clin Endocrinol Metab. 2004;89: [PMID: ] 8. Douma S, Petidis K, Doumas M, Papaefthimiou P, Triantafyllou A, Kartali N, et al. Prevalence of primary hyperaldosteronism in resistant hypertension: a retrospective observational study. Lancet. 2008;371: [PMID: ] 9. Melby JC, Spark RF, Dale SL, Egdahl RH, Kahn PC. Diagnosis and localization of aldosterone-producing adenomas by adrenal-vein cateterization. N Engl J Med. 1967;277: [PMID: ] 10. Doppman JL, Gill JR Jr, Miller DL, Chang R, Gupta R, Friedman TC, et al. Distinction between hyperaldosteronism due to bilateral hyperplasia and unilateral aldosteronoma: reliability of CT. Radiology. 1992;184: [PMID: ] 11. Magill SB, Raff H, Shaker JL, Brickner RC, Knechtges TE, Kehoe ME, et al. Comparison of adrenal vein sampling and computed tomography in the differentiation of primary aldosteronism. J Clin Endocrinol Metab. 2001;86: [PMID: ] 12. Harper R, Ferrett CG, McKnight JA, McIlrath EM, Russell CF, Sheridan B, et al. Accuracy of CT scanning and adrenal vein sampling in the pre-operative localization of aldosterone-secreting adrenal adenomas. QJM. 1999;92: [PMID: ] 13. McAlister FA, Lewanczuk RZ. Primary hyperaldosteronism and adrenal incidentaloma: an argument for physiologic testing before adrenalectomy. Can J Surg. 1998;41: [PMID: ] 14. Freriks K, Schultze Kool LJ, Timmers HJ, Deinum J, Lenders JW, Hermus AR. [Determining aldosterone in the adrenal veins in order to establish uni- or bilateral aldosterone production in patients with primary hyperaldosteronism]. Ned Tijdschr Geneeskd. 2007;151: [PMID: ] 15. Young WF, Stanson AW, Thompson GB, Grant CS, Farley DR, van Heerden JA. Role for adrenal venous sampling in primary aldosteronism. Surgery. 2004;136: [PMID: ] 16. Funder JW, Carey RM, Fardella C, Gomez-Sanchez CE, Mantero F, Stowasser M, et al. Case detection, diagnosis, and treatment of patients with primary aldosteronism: an endocrine society clinical practice guideline. J Clin Endocrinol Metab. 2008; 93: Rossi GP. Surgically correctable hypertension caused by primary aldosteronism. Best Pract Res Clin Endocrinol Metab. 2006;20: [PMID: ] 18. Doppman JL, Gill JR Jr. Hyperaldosteronism: sampling the adrenal veins. Radiology. 1996;198: [PMID: ] 19. Ishibashi T, Satoh F, Yamada T, Sato A, Matsuhashi T, Takase K. Primary aldosteronism: a pictorial essay. Abdom Imaging. 2007;32: [PMID: ] 20. Nwariaku FE, Miller BS, Auchus R, Holt S, Watumull L, Dolmatch B, et al. Primary hyperaldosteronism: effect of adrenal vein sampling on surgical outcome. Arch Surg. 2006;141: ; discussion [PMID: ] 21. Stowasser M, Gordon RD. Primary aldosteronism careful investigation is essential and rewarding. Mol Cell Endocrinol. 2004;217:33-9. [PMID: ] 22. Bernini G, Moretti A, Argenio G, Salvetti A. Primary aldosteronism in normokalemic patients with adrenal incidentalomas. Eur J Endocrinol. 2002; 146: [PMID: ] 23. Castro OL, Yu X, Kem DC. Diagnostic value of the post-captopril test in primary aldosteronism. Hypertension. 2002;39: [PMID: ] 24. Geisinger MA, Zelch MG, Bravo EL, Risius BF, O Donovan PB, Borkowski GP. Primary hyperaldosteronism: comparison of CT, adrenal venography, and venous sampling. AJR Am J Roentgenol. 1983;141: [PMID: ] 25. Mulatero P, Bertello C, Rossato D, Mengozzi G, Milan A, Garrone C, et al. Roles of clinical criteria, computed tomography scan, and adrenal vein sampling in differential diagnosis of primary aldosteronism subtypes. J Clin Endocrinol Metab. 2008;93: [PMID: ] 26. Hambling C, Jung RT, Browning MC, Gunn A, Anderson JM. Primary hyperaldosteronism evaluation of procedures for diagnosis and localization. Q J Med. 1993;86: [PMID: ] 27. Gleason PE, Weinberger MH, Pratt JH, Bihrle R, Dugan J, Eller D, et al. Evaluation of diagnostic tests in the differential diagnosis of primary aldosteronism: unilateral adenoma versus bilateral micronodular hyperplasia. J Urol. 1993; 150: [PMID: ] 28. Linde R, Coulam C, Battino R, Rhamy R, Gerlock J, Hollifield J. Localization of aldosterone-producing adenoma by computed tomography. J Clin Endocrinol Metab. 1979;49: [PMID: ] 29. Toniato A, Bernante P, Rossi GP, Pelizzo MR. The role of adrenal venous sampling in the surgical management of primary aldosteronism. World J Surg. 2006;30: [PMID: ] 30. Harvey A, Kline G, Pasieka JL. Adrenal venous sampling in primary hyper September 2009 Annals of Internal Medicine Volume 151 Number 5

9 Diagnostic Procedures to Differentiate Adrenal Abnormalities Review aldosteronism: comparison of radiographic with biochemical success and the clinical decision-making with less than ideal testing. Surgery. 2006;140:847-53; discussion [PMID: ] 31. Omura M, Sasano H, Saito J, Yamaguchi K, Kakuta Y, Nishikawa T. Clinical characteristics of aldosterone-producing microadenoma, macroadenoma, and idiopathic hyperaldosteronism in 93 patients with primary aldosteronism. Hypertens Res. 2006;29: [PMID: ] 32. Carr CE, Cope C, Cohen DL, Fraker DL, Trerotola SO. Comparison of sequential versus simultaneous methods of adrenal venous sampling. J Vasc Interv Radiol. 2004;15: [PMID: ] 33. Glodny B, Kühle C, Cromme S, Brockmann J, Winde G. An assessment of diagnostic procedures preparatory to retroperitoneoscopic removal of adenoma in cases of primary hyperaldosteronism. Endocr J. 2000;47: [PMID: ] 34. Sheaves R, Goldin J, Reznek RH, Chew SL, Dacie JE, Lowe DG, et al. Relative value of computed tomography scanning and venous sampling in establishing the cause of primary hyperaldosteronism. Eur J Endocrinol. 1996;134: [PMID: ] 35. Blumenfeld JD, Sealey JE, Schlussel Y, Vaughan ED Jr, Sos TA, Atlas SA, et al. Diagnosis and treatment of primary hyperaldosteronism. Ann Intern Med. 1994;121: [PMID: ] 36. White ML, Gauger PG, Doherty GM, Cho KJ, Thompson NW, Hammer GD, et al. The role of radiologic studies in the evaluation and management of primary hyperaldosteronism. Surgery. 2008;144:926-33; discussion 933. [PMID: ] 37. Schwab CW 2nd, Vingan H, Fabrizio MD. Usefulness of adrenal vein sampling in the evaluation of aldosteronism. J Endourol. 2008;22: [PMID: ] 38. Minami I, Yoshimoto T, Hirono Y, Izumiyama H, Doi M, Hirata Y. Diagnostic accuracy of adrenal venous sampling in comparison with other parameters in primary aldosteronism. Endocr J. 2008;55: [PMID: ] 39. Satoh F, Abe T, Tanemoto M, Nakamura M, Abe M, Uruno A, et al. Localization of aldosterone-producing adrenocortical adenomas: significance of adrenal venous sampling. Hypertens Res. 2007;30: [PMID: ] 40. Wu KD, Liao TS, Chen YM, Lai MK, Chen SJ, Su CT, et al. Preoperative diagnosis and localization of aldosterone-producing adenoma by adrenal venous sampling after administration of metoclopramide. J Formos Med Assoc. 2001; 100: [PMID: ] 41. Wu VC, Chueh SC, Chang HW, Lin WC, Liu KL, Li HY, et al. Bilateral aldosterone-producing adenomas: differentiation from bilateral adrenal hyperplasia. QJM. 2008;101: [PMID: ] 42. Wang JH, Wu HM, Sheu MH, Tseng HS, Chiang JH, Chang CY. High resolution MRI of adrenal glands in patients with primary aldosteronism. Zhonghua Yi Xue Za Zhi (Taipei). 2000;63: [PMID: ] 43. Lo CY, Tam PC, Kung AW, Lam KS, Wong J. Primary aldosteronism. Results of surgical treatment. Ann Surg. 1996;224: [PMID: ] 44. Rao A, Melby JC, Wilson TE. Prohormones in adrenal venous effluent in patients with primary hyperaldosteronism. J Clin Endocrinol Metab. 1995;80: [PMID: ] 45. Naruse K, Naruse M, Tanabe A, Yoshimoto T, Watanabe Y, Kurimoto F, et al. Does plasma immunoreactive ouabain originate from the adrenal gland? Hypertension. 1994;23:I [PMID: ] 46. Dunnick NR, Leight GS Jr, Roubidoux MA, Leder RA, Paulson E, Kurylo L. CT in the diagnosis of primary aldosteronism: sensitivity in 29 patients. AJR Am J Roentgenol. 1993;160: [PMID: ] 47. Takasaki I, Shionoiri H, Yasuda G, Miyajima E, Umemura S, Gotoh E, et al. Preoperative lateralisation of aldosteronomas by aldosterone/cortisol ratios in adrenal venous plasma. J Hum Hypertens. 1987;1:95-9. [PMID: ] 48. Ma JT, Wang C, Lam KS, Yeung RT, Chan FL, Boey J, et al. Fifty cases of primary hyperaldosteronism in Hong Kong Chinese with a high frequency of periodic paralysis. Evaluation of techniques for tumour localisation. Q J Med. 1986;61: [PMID: ] 49. Witzgall H, Müller OA, Weber PC. Clinical and biochemical features of patients with aldosterone-producing adenoma and idiopathic hyperaldosteronism. Klin Wochenschr. 1983;61: [PMID: ] 50. Dunnick NR, Doppman JL, Gill JR Jr, Strott CA, Keiser HR, Brennan MF. Localization of functional adrenal tumors by computed tomography and venous sampling. Radiology. 1982;142: [PMID: ] 51. McAreavey D, Brown JJ, Cumming AM, Davidson JK, Duncan JG, Fraser R, et al. Pre-operative localization of aldosterone-secreting adrenal adenomas. Clin Endocrinol (Oxf). 1981;15: [PMID: ] 52. Nocaudie-Calzada M, Huglo D, Lambert M, Ernst O, Proye C, Wemeau JL, et al. Efficacy of iodine-131 6beta-methyl-iodo-19-norcholesterol scintigraphy and computed tomography in patients with primary aldosteronism. Eur J Nucl Med. 1999;26: [PMID: ] 53. Volpe C, Enberg U, Sjögren A, Wahrenberg H, Jacobsson H, Törring O, et al. The role of adrenal scintigraphy in the preoperative management of primary aldosteronism. Scand J Surg. 2008;97: [PMID: ] 54. Rossi GP, Ganzaroli C, Miotto D, De Toni R, Palumbo G, Feltrin GP, et al. Dynamic testing with high-dose adrenocorticotrophic hormone does not improve lateralization of aldosterone oversecretion in primary aldosteronism patients. J Hypertens. 2006;24: [PMID: ] 55. Espiner EA, Ross DG, Yandle TG, Richards AM, Hunt PJ. Predicting surgically remedial primary aldosteronism: role of adrenal scanning, posture testing, and adrenal vein sampling. J Clin Endocrinol Metab. 2003;88: [PMID: ] 56. Kline GA, Harvey A, Jones C, Hill MH, So B, Scott-Douglas N, et al. Adrenal vein sampling may not be a gold-standard diagnostic test in primary aldosteronism: final diagnosis depends upon which interpretation rule is used. Variable interpretation of adrenal vein sampling. Int Urol Nephrol. 2008;40: [PMID: ] 1 September 2009 Annals of Internal Medicine Volume 151 Number 5 337

10 Annals of Internal Medicine Current Author Addresses: Drs. Kempers, Lenders, van der Wilt, Schultze Kool, Hermus, and Deinum and Ms. Outheusden: Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, the Netherlands. W September 2009 Annals of Internal Medicine Volume 151 Number 5