Clinical Biochemistry 46 (2013) 1399–1404

Contents lists available at ScienceDirect

Clinical Biochemistry

j ourna l homepage: www.e lsev ie r .com/ locate /c l inb iochem

Adrenal vein sampling: Substantial need for technical improvement at regionalreferral centres

Panda Elliott a, Daniel T. Holmes b,c,⁎a Faculty of Arts and Science, Queen's University, Kingston ON, Canadab University of British Columbia Department of Pathology and Laboratory Medicine, Vancouver, BC, Canadac St. Paul's Hospital Department of Pathology and Laboratory Medicine, Vancouver, BC, Canada

Abbreviations: AVS, Adrenal Vein Sampling; ACTH, AAPA, Aldosterone Producing Adenoma; BAH, Bilateral AColumbia; IVC, Inferior Vena Cava; LAV, Left Adrenal VeLifeLabs; PA, Primary Aldosteronism; PRA, Plasma ReniMass; RAV, Right Adrenal Vein; SAC, Serum AldosteronIndex; SPH, St. Paul's Hospital.⁎ Corresponding author at: Department of Pathology an

Hospital, 1081 Burrard St, Vancouver, BC, V6Z 1Y6 CanadaE-mail address: dtholmes@mail.ubc.ca (D.T. Holmes

0009-9120/$ – see front matter © 2013 The Canadian Shttp://dx.doi.org/10.1016/j.clinbiochem.2013.04.004

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 17 January 2013Received in revised form 12 March 2013Accepted 7 April 2013Available online 16 April 2013

Keywords:Primary aldosteronismConn's syndromeAdrenal vein samplingMineralocorticoidAdrenal hypertension

Object: Adrenal vein sampling (AVS) is the gold standard for localization of aldosterone producing adenoma.The anatomy of the right adrenal vein makes this procedure technically demanding and it may yield no clinicalinformation if the adrenal veins are not adequately cannulated. Having frequently observed the technical failureof AVS, we undertook a review of 220 procedures in British Columbia, Canada.

Design and methods: Subjects were retrospectively identified through the laboratory information system.The following were collected: demographics, screening aldosterone concentration and renin activity/mass,results of dynamic function tests, AVS aldosterone and cortisol results. Standard calculations were performedon AVS data and site-specific success rates were compared. The effect of adrenocorticotropin hormone (ACTH)stimulation on the selectivity index (SI) and lateralization index (LI) were explored.

Results: The overall technical success-rate of AVS procedures was only 44% in procedures where noACTH-stimulation was used (n = 200) but this rose significantly (p b 0.01) to 82% for those employing ACTH

(n = 139). ACTH-stimulation significantly increased the median SI (left: 5.8 vs 36.7, p b 0.01; right: 7.0 vs51.2, p b 0.01), and salvaged 36 procedures from yielding no information, 21 of which demonstrated lateraliza-tion of aldosterone production. In 64 cases showing lateralization both pre and post-stimulation, ACTH signifi-cantly decreased themedian LI from 5.4 to 2.2, p b 0.01, creating substantial risk for spurious loss of lateralization.

Conclusions: The technical success of AVS is lower than reported elsewhere. Provided that effects on the LIare considered, the use of ACTH-stimulation during AVS assists in the identification of unilateral forms of PA.

© 2013 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.

Introduction

Epidemiology

The prevalence of hypertension among adults in Canada and the UShas been estimated at 22–23%with future increases forecast [1]. Conse-quences of untreated hypertension include: higher risk of coronary andperipheral vascular disease, left ventricular hypertrophy, congestiveheart failure, and chronic renal impairment. Accordingly, identificationand treatment of secondary hypertension is imperative to chronicdisease prevention.

drenocorticotropin Hormone;drenal Hyperplasia; BC, Britishin; LI, Lateralization Index; LL,n Activity; PRM, Plasma Renine Concentration; SI, Selectivity

d LaboratoryMedicine, St. Paul's. Fax: +1 604 806 8815.).

ociety of Clinical Chemists. Publishe

Clinical presentation

Approximately 11% of newly diagnosed hypertension is caused byprimary aldosteronism (PA) [2]. PA is characterized by inappropriateadrenal production of aldosterone in the absence of renin-mediatedstimulation [3]. Patients typically present with high-normal to overtlyhigh serum aldosterone concentration (SAC), and low-normal toundetectable plasma renin activity (PRA) or plasma renin mass(PRM), treatment-resistant hypertension, and may also have hypoka-lemia and metabolic alkalosis. It is noteworthy that most patients arenormokalemic [4,5].

Screening and confirmation

Biochemical screening for PA is undertaken in hypertensive individ-uals displaying any of the following: moderate to severe hypertension,early age of onset, treatment-resistance, concomitant spontaneous ordiuretic-induced hypokalemia, or documented adrenal incidentaloma[3–6]. Screening is performed by analysis of SAC and PRA (or PRM)after appropriate patient preparation [3–6]. A positive screen for PA isconstituted by a SAC:PRA ratio above aminimum threshold determined

d by Elsevier Inc. All rights reserved.

1400 P. Elliott, D.T. Holmes / Clinical Biochemistry 46 (2013) 1399–1404

by the properties of the assays employed [5,6]. It has additionallybeen suggested that SAC should be above a minimum concentrationbefore the SAC:PRA ratio is considered because suppressed PRA canlead to spuriously positive screens even in low-aldosterone states [3–5].Diagnostic confirmation of PA is accomplished using one of several dy-namic function tests: saline infusion, oral salt loading, fludrocortisonesuppression, or captopril suppression followed by SAC or 24-h urinecollections as appropriate [5–7].

Adrenal vein sampling

After diagnostic confirmation of PA, subtype classification isrequired to identify unilateral (surgically-curable) forms, most oftenaldosterone producing adenomas (APAs) [3–6], which represent 30–35% of cases [3,4]. Use of computed tomography imaging for subtypeclassification is unreliable because many APAs are too small to bedetected and non-functioning incidentalomas may be mistakenlyidentified as causative. Therefore, the gold-standard for subtype clas-sification is direct sampling and analysis of aldosterone and cortisolfrom adrenal vein effluent bilaterally, commonly referred to as adre-nal vein sampling (AVS).

The AVS procedure is performed under fluoroscopic guidance by aninterventional radiologist [8] and is technically demanding due to theanatomy of the right adrenal vein (RAV), which inserts directly intothe posterior aspect of inferior vena cava (IVC) in proximity to a numberof other veins of similar calibre. This canmake anatomical identificationof the RAV difficult. Furthermore, the RAV is usually shorter andnarrower than the left adrenal vein (LAV) and may occasionally entera common drainage with a small hepatic or renal capsular accessoryveins [8]. It is important to avoid the incidental collection of hepaticdrainage because it is aldosterone-deplete and may thereby confoundcalculations [9]. In contrast, the LAV drains into the superior aspect ofthe left renal vein, making it easier to locate. In an AVS procedure, theRAV and LAV are respectively or concomitantly catheterized, uponwhich samples from both are collected and analyzed for cortisol andaldosterone. It is important to collect samples from the RAV before theLAV because the former takes longer to identify and this order of collec-tion helps ensure samples are drawn in close temporal proximity. Thismitigates the risk of ACTH-mediated (stress-induced) fluctuations inthe aldosterone/cortisol secretion which may confound interpretation.

Cortisol results from putative adrenal vein candidates are dividedby those of the peripheral blood (lower-IVC, iliac or femoral vein) tocalculate the so-called “selectivity index” (SI — see Table 1). The SImust exceed a pre-determined threshold in order to prove successfulcannulation and appropriate sampling of adrenal effluent. An SI ≥2 is

Table 1Definitions for the selectivity index, lateralization index, and thresholds for their interpre-tationwhichwere employed in this study. “Periphery”, in reference to the SI calculation, isunderstood tomean a collection from the infrarenal inferior vena cava, the iliac vein or thefemoral vein.

Term Definition

Selectivity index (SI) Cort(adrenal)/Cort(periphery)Successful catheterization pre-ACTHstimulation

SI ≥ 2

Successful catheterization post-ACTHstimulation

SI ≥ 3

Unsuccessful catheterization on the right(USCR)

SI b 2 (pre-ACTH)OR SI b 3 (post-ACTH)

Unsuccessful catheterization on the left(USCL)

SI b 2 (pre-ACTH)OR SI b 3 (post-ACTH)

Bilaterally unsuccessful catheterization(BUSC)

USCR AND USCL

Lateralization index (LI) [A/C(dominant)]/[A/C(non-dominant)]Aldosterone producing adenoma(LAA/RAA)

LI ≥ 4

Indeterminate diagnosis (IND) 3 b LI b 4Bilateral adrenal hyperplasia (BAH) LI ≤ 3

often considered evidence of adequate cannulation in the absenceof ACTH-stimulation while levels ≥3 are used for AVS proceduresaugmented with ACTH [10]. If cannulation is successful for bothadrenal veins, further calculations can be undertaken. In order to deter-mine whether the cause of PA is unilateral (and therefore surgicallytreatable), the lateralization index (LI) is calculated dividing the larger(“dominant”) venous aldosterone-to-cortisol (A/C) ratio by the smaller(“non-dominant”) one (see Table 1).

There is no consensus on a minimum LI to diagnose lateralization[5–10]. Results from 3 to 5 are often considered specific to unilateralforms of PA, but use of higher values in this range increases the spec-ificity for unilateral disease at the cost of diagnostic accuracy [11,12].Results below 3 are often used to define bilateral disease and someauthors propose a “gray” or indeterminate zone for results between3 and 4 [10].

As mentioned, the stressful conditions placed upon patients duringAVS are thought to be a risk for sporadic intraprocedural aldosterone/cortisol secretion, potentially confounding calculations. To overcomethis effect, some sites use ACTH-infusion or bolus to overwhelm anyphysiological effects [5]. The use of ACTH also exaggerates cortisolsecretion making biochemically-proven cannulation easier to attain[11].

Screening in British Columbia Canada

In British Columbia (BC — population 4.6 million), PA screening isperformed at two laboratories: LifeLabs (LL) in Victoria and St. Paul'sHospital (SPH) in Vancouver. The two laboratories use differentapproaches for PA screening: LL uses Siemens Coat-a-Count® radioim-munoassay for SAC and the Diasorin Liaison chemiluminescent sand-wich assay for PRM. SPH uses liquid chromatography and tandemmass spectrometry for aldosterone [13] and an in-house radioimmuno-assay (RIA) for PRA [14]. Notwithstanding the difference in renin deter-mination, both strategies are known to be effective for PA screening[15].

Analysis of all AVS samples in BC occurs at SPH. When SPH proto-colized the analysis and interpretation of AVS in 2006, it was apparentthat rates of bilaterally successful cannulation of the adrenal veinswere much lower than the >95% reported elsewhere [3]. Furthermore,there was disparity in success-rates between radiological operators.Because AVS exposes the patient to certain risks (infection, radiationexposure and rarely: adrenal vein rupture/thrombosis, or adrenal infarc-tion) and because the procedure is costly from the perspective ofconsumables, radiologist/nursing time, use of the fluoroscopy suite,biochemical analysis, and interpretive time [16], a review of provincialAVS success was warranted as a quality-improvement initiative.

We have undertaken a 10 year retrospective review of AVS proce-dures in BC to investigate the site-specific success rates and to identifythe practices of more successful sites.

Materials methods

Method of data collection

This study was approved by the SPH ethics board. One hundred andninety-eight subjects who underwent 220 AVS procedures in BC over a10-year period were identified through the laboratory informationsystem at SPH. Extracted demographic information collected included:gender, age of initial screen, and age at AVS. Extracted biochemicaldata collected included: aldosterone and PRA/PRM results at the timeof first screening, the dynamic function test employed (saline suppres-sion or oral salt load) including the accompanying aldosterone results(serum/urine), and all collection dates. With respect to AVS, the follow-ing were extracted: hospital performing AVS and date of procedure,aldosterone and cortisol results for the LAV, RAV and peripheral vein.Results from all 6 institutions performing AVS (herein denoted A–F)

1401P. Elliott, D.T. Holmes / Clinical Biochemistry 46 (2013) 1399–1404

were reviewed. Collection of AVS samples varied by site, with somesites collecting samples without ACTH stimulation (A,B, and F), oneusing continuous ACTH infusion followed by bolus (D), and some sitescollecting before and after high dose IV (250 μg) ACTH administration(C and E). Where ACTH was employed, aldosterone and cortisol resultsfor the LAV, RAV and peripheral blood were obtained both pre andpost-ACTH.

Biochemical analysis

All analyses were performed according to the manufacturer'srecommendation. For the subset of patients screened at LL, SAC wasperformed using the assays described above (SAC by SiemensCoat-a-Count® RIA, PRM by Diasorin Liaison®). For the patientsscreened at SPH, three different methods were used for aldosteroneover a 10-year period: Adaltis MAIA RIA (until July 2007), SiemensCoat-a-Count® RIA (July 2007–January 2011), and tandem mass spec-trometry (January 2011 forward) [13]. Because method-dependentbiases are problematic for aldosterone [17,18], results from differentmanufacturers cannot be compared without normalization. Therefore,for the purposes of statistical analysis in this study, all screening andconfirmatory aldosterone data were normalized by linear transforma-tion to the Siemens Coat-a-Count® RIA using appropriate regressioncoefficients. The regression relationship (Passing Bablok) betweenLC–MS/MS and the Siemens Coat-a-Count® RIA was established on acohort of 118 specimens taken from 111 unique subjects as previouslydescribed [13]. The regression equation was LC–MS/MS = 1.17 ×Siemens-63.9 pmol/L, CI slope [1.08 to 1.25]; CI intercept [−81.5 to−47.9]. The regression relationship between the Adaltis MAIA RIAand Siemens Coat-a-Count® RIA was established on a cohort of 36subjects from routine outpatient work. The regression equationwas Adaltis =1.61 × Siemens + 76.2 pmol/L, CI slope [1.48 to 1.74],CI intercept [37.5 to 95.1]. At SPH, PRA was evaluated by the samein-house RIA over the entire period of the study [14].

All AVS SAC analysis occurred at SPH using the assays described.Aldosterone was analyzed on the RAV, LAV at three dilutions: 1:1,1:40, and 1:80 (for RIA methods) or at 1:50 and 1:100 (for mass spec-trometry). Of the 3 results obtained for each adrenal sample, the onewhose diluted value was closest to the midpoint of the analyticalrange was reported. AVS cortisol was analyzed on the SiemensImmulite 2000 (until June 2007), the Siemens Immulite 2500 (June2007–January 2011) and simultaneously by mass spectrometry with

Table 2Demographic information and results of screening and dynamic function studies. To convertstudy generates angiotensin I at pH = 7.4 instead of pH = 6.0, as with most commercial PRof commercial methods.

Male

Age(y)Median (IQR) 54 (48–58)Range 28–75

Screening SAC (pmol/L), n = 173Median (IQR) 610 (447–830)Range 231–2148

Screening PRA (ng/L/s), n = 150Median b0.05Range b0.05–0.46

Screening PRM (ng/L), n = 25a

Median (IQR) 2.1 (2.0–3.3)Range 2.0–6.9

Post SST SAC (pmol/L), n = 75Median (IQR) 286 (187–418)Range 48–1183

Confirmation urinaryAldosterone excretion (nmol/d), n = 16

Median (IQR) 81 (64–159)Range 49–373

a Does not include patients who were screened at LL prior to being screened at SPH.

aldosterone (January 2011 and forward). For the Immulite instruments,necessary dilutionswere handled by on-board algorithms, and dilutionsfor mass spectrometry were 1:1, 1:50, and 1:100. Because the LIunitless, normalization of AVS aldosterone results was unnecessary.

Calculations and interpretation

Until early 2011 screening tests performed at SPH were consideredpositive for PA if the absolute aldosterone concentration was greaterthan 415 pmol/L (15 ng/dL) and a minimum SAC:PRA of 2500 pmol/Lper ng/L/swas observed. Aftermigration tomass-spectrometric analysis,the aldosterone threshold was lowered to 300 pmol/L (10.8 ng/dL) onthe a basis of published evidence [19]; no change to SAC:PRA thresholdwas implemented. Screening results from LL were considered positiveif the SAC:PRM exceeded 150 pmol/L per ng/L (5.4 ng/dL per ng/L)until June 2010, after which this was lowered to 100 pmol/L per ng/L(3.6 ng/dL per ng/L). For dynamic function testing, in our practice,post-saline aldosterone results above 138 pmol/L (5 ng/dL) are consid-ered diagnostic for PA as are 24-h urine aldosterone excretions greaterthan 38.8 nmol/d (14 μg/d) after oral salt load, provided urinary sodiumexcretion exceeds 200 mmol/d.

Based on the cortisol results from the RAV, LAV and peripheralveins, the right and left SI were each calculated to determine if theprocedure was technically successful. Technical success was definedas bilateral adrenal–venous to peripheral–venous cortisol ratio of≥2 for unstimulated collections, and ≥3 for a ACTH-stimulated pro-cedure in accordance with previously published thresholds [10].When the adrenal vein cannulation was bilaterally successful, the LIwas calculated and interpreted using the definitions described inTable 1. All statistical analysis was performed using R version 2.13.1.

Results

Two-hundred and twenty AVS procedures were performed on 198unique subjects (113 male, 85 female). Demographic data, screeningaldosterone and PRA/PRM results and results of available dynamicfunction testing are shown in Table 2.

Eighty-one procedures (from sites A, B, E, and F) were performedwithout any ACTH stimulation, while 119 (from sites C and E) wereperformed with samples collected both pre and post high-doseACTH stimulation. The remaining 20 (site D) were drawn during an

aldosterone expressed in pmol/L to ng/dL, divide by 27.7. The PRA method used in thisA assays. The effect of this pH difference is that numerical results are approximately 1/2

Female All

48 (40–60) 53 (45–58)22–69 22–75

588 (400–810) 592 (430–828)161–3093 161–3093

b0.05 b0.05b0.05–0.43 b0.05–0.46

2.6 (2.0–4.5) 2.1 (2.0–4.4)2.0–11.7 2.0–11.7

371 (191–688) 291 (187–530)63–2118 48–2118

56 (46–69) 65 (48–133)38–380 38–380

Table 3Results and success rates of AVS procedures by site. Successes were defined as collec-tions for which the SI for both LAV and RAV candidates was ≥2 for samples collectedwithout ACTH stimulation and greater than ≥3 for those collected after ACTH stimula-tion. See Table 1 for diagnostic abbreviations.

Centre A B C D E F Total

Number of radiologists ≥1 ≥1 1 2 7 7 ≥19Cases per operator ≤3 ≤4 87 10 12.1 3

Collections without ACTH stimulationPatients receiving AVS 3 4 87 NA 85 21 200

USCL 0 1 14 NA 9 0 24 (12%)USCR 1 1 15 NA 18 9 44 (22%)BUSC 2 2 9 NA 20 7 40 (20%)IND 0 0 4 NA 1 0 5 (2.5%)LAA 0 0 13 NA 10 0 23 (11.5%)RAA 0 0 13 NA 8 4 25 (12.5%)BAH 0 0 18 NA 16 1 35 (17.5%)Incomplete collections 0 0 1 NA 3 0 4 (2%)

Success rate (%) 0 0 55 NA 41 24 44

Collections with ACTH stimulationPatients receiving AVS NA NA 87 20 32 NA 139

USCL NA NA 0 0 1 NA 1 (1%)USCR NA NA 9 6 7 NA 22 (16%)BUSC NA NA 0 0 1 NA 1 (1%)IND NA NA 5 3 0 NA 8 (6%)LAA NA NA 19 4 5 NA 28 (20%)RAA NA NA 16 5 4 NA 25 (18%)BAH NA NA 38 2 13 NA 53 (38%)Incomplete collections 0 0 0 0 1 0 0

Success rate (%) NA NA 90 70 69 NA 82

1402 P. Elliott, D.T. Holmes / Clinical Biochemistry 46 (2013) 1399–1404

ACTH infusion (25 μg/h for prior 30 min) and following a subsequentbolus (15 μg).

The cannulation success rates of AVS testing by site along with therates of lateralization and bilateral adrenal hyperplasia (BAH) areshown in Table 3. At least 19 radiologists performed AVS over allsites and the number of operators by site ranged from 1 to 7. Theoverall technical success-rate of the procedure was 44% forunstimulated collections and 82% for ACTH-stimulated collections(88/200 vs 114/139, p b 0.01, χ2 test). By site, success-rate rangedfrom 0% (sites A and B: 0/3 and 0/4 cases) to 55% (site C: 48/87cases) for unstimulated collections and 69% (22/32 cases) to 90%(78/87 cases) for ACTH-stimulated collections. Of the 90 casesdemonstrating successful cannulation without ACTH stimulation, 48(53%) showed lateralization, 35 (39%) were bilateral and 5 (5.6%)were indeterminate. For the 114 successful cannulations performed

Fig. 1. A: Boxplots of the effect of ACTH stimulation on the SI in cases where successful caB: Boxplot of the effect of ACTH stimulation on the LI of AVS procedures showing successleft or right) in the pre-ACTH results.

with ACTH-stimulation, 53 (46%) cases lateralized, 53 (46%) werebilateral, and 8 (7%) were indeterminate.

Effect of ACTH on cannulation success and selectivity index

For those AVS procedures in which samples were collected bothpre and post-ACTH stimulation (n = 117, with complete data, fromsites C and E), the overall pooled technical success rate higher afterACTH stimulation (47.8% vs 85.5% of cases, Fisher's Exact Test,p b 0.01). To more specifically compare the effects of ACTH, welooked at cases showing bilaterally successful cannulation in boththe pre and post-ACTH phases (n = 64, from sites C and E). In thissubset, ACTH significantly increased the median SI: pre-L SI: 5.8, IQR[2.75–13.2] vs post-L SI: 36.7, IQR [21.1–55.8], p b 0.01 Wilcoxonsigned rank; pre-R SI: 7.0 [3.5–20.8] vs post-R SI: 51.2 [38.2–64.2],p b 0.01, Wilcoxon signed rank, as shown in Fig. 1A.

Effect of ACTH on lateralization index

Considering these same cases bilaterally cannulated in both thepre and post-ACTH phase, ACTH-stimulation caused a statistically sig-nificant drop in the median LI from pre: 5.4, IQR [1.8–15.9] to post:2.2, IQR [1.4–7.6], p b 0.01, Wilcoxon signed rank, as shown inFig. 1B. More importantly, this potentially deleterious effect extendedto those 44 cases showing lateralization in the pre-ACTH collection.The median LI dropped for these patients dropped significantly frompre:13.4, IQR [8.7–29.3] to post: 6.1 IQR [2.3–12.3], p b 0.01,Wilcoxon signed rank. Not surprisingly, in 14 of 44 of these casesthe interpretation changed in the post-ACTH phase to indeterminate(3 b LI b 4, n = 1) or bilateral (LI ≤ 3, n = 13).

Effect of hospital site

There were statistically significant differences in technical successrates between sites for collections performed without ACTH-stimulation(taken from A–C, E–F) (Fisher's Exact Test, p b 0.01) and between sitesC–E for ACTH-stimulated collections (Fisher's Exact Test, p b 0.01).Though site C appeared superior, after accounting for multiple pairwisecomparisons [20], this higher success-rate is statistically significant inthe post-ACTH phase only (Fisher's Exact Test, p = 0.03 between sitesC and D, p = 0.05 between sites C and E).

nnulation was achieved in the pre and post-ACTH phases of the collection (n = 64).ful cannulation both pre and post-ACTH stimulation and lateralization (LI > 4, either

1403P. Elliott, D.T. Holmes / Clinical Biochemistry 46 (2013) 1399–1404

Is technical success better in more recent collections?

Limiting calculations to more recent collections (January 2010 andforward) allows assessment of improvement over time; results areshown in Table 4. Overall province-wide technical success-rates areonly modestly different: 47% (vs. 44%, p = 0.72, χ2 test) in casesperformed without ACTH stimulation (n = 87) and 86% (vs. 82%,p = 0.52, χ2 test) in ACTH-stimulated cases (n = 87). In thistime-frame, success rates between sites (A, C–F) remained signifi-cantly different in unstimulated collections (Fisher's Exact Test,p = 0.02) but differences did not reach statistical significance inpairwise comparison. Success-rates in ACTH-stimulated collectionsamong sites C–E remained significantly different (Fisher's ExactTest, p b 0.01) and the superior performance of site C (55/56 cases,98.2%) was again statistically significant compared with D and Erespectively (both p b 0.01).

Discussion

Results of this retrospective review demonstrate that AVS may notbe as successful in real-world practice as reported by specializedcentres [21–23]. However, others have observed similarly low cannula-tion success rates to those shown here. Vonend et al. recently publisheda case-series from the German Conn-Syndrome Registry retrospectivelyand prospectively examining results of 200 AVS procedures performedat 5 centres without ACTH-stimulation [24]. This data demonstratedthat only 31% of cases were technically successful using an SI thresholdof 2.0. It was also observed that sites performing b20 procedures hadsuccess rates between 8 and 10%.

However, it should be pointed out that not all of the proceduresappearing unsuccessful were necessarily failed cannulations. The SIthresholds adopted define the cannulation success andmore permissivethresholds obviously result in higher success rates [24]. Some centresuse SI thresholds as low as 1.10 [25,26]. The hazard of using sucha low threshold is that cortisol assays have within-run analytical(im)precision (CV) in the range of ~3–7% [27]. Ignoring biologicaleffects, which may be substantial, to declare a significant differencebetween serial analysis of samples with the same concentration, resultswould have differ by more than 2.8 standard deviations or ~8–20%. For

Table 4Results and success rates of AVS procedures by site for collections performed January2010 and forward. See caption in Table 1 for diagnostic abbreviations.

Centre A B C D E F Total

Number of radiologists NA 0 1 2 2 4 9Cases per operator 1 0 56 5 10.5 2.3

Collections without ACTH stimulationPatients receiving AVS 1 0 56 NA 21 9 87

USCL 0 0 11 NA 1 0 12 (13.8%)USCR 1 0 6 NA 3 4 14 (16.1%)BUSC 0 0 5 NA 5 4 14 (16.1%)IND 0 0 4 NA 0 0 4 (4.6%)LAA 0 0 7 NA 3 0 10 (11.5%)RAA 0 0 9 NA 3 1 13 (14.9%)BAH 0 0 14 NA 6 0 20 (23%)Incomplete collections 0 0 0 NA 0 0 0

Success rate (%) 0 0 61 NA 57 11 47

Collections with ACTH stimulationPatients receiving AVS NA NA 56 10 21 NA 87

USCL NA NA 0 0 1 NA 1 (1%)USCR NA NA 1 4 6 NA 22 (16%)BUSC NA NA 0 0 0 NA 1 (1%)IND NA NA 2 0 0 NA 8 (6%)LAA NA NA 11 3 3 NA 28 (20%)RAA NA NA 15 3 1 NA 25 (18%)BAH NA NA 27 0 10 NA 53 (38%)Incomplete collections 0 0 0 0 0 0 0

Success rate (%) NA 98 60 67 NA 86

this reason, a 10% increase may be a random effect. Not surprisingly,protocols with highly permissive SI demonstrate poor reproducibility[28].

Evidently, there are occasionally compelling reasons to relax SIcriteria on a case-by-case basis. For example, Table 5 highlights theAVS results of a 53 year old male included in this study who had PAand a 1 cm right-sided adrenal adenoma on CT scan. Though theright-sided SI values of 1.47 and 1.51 do not support successful cannu-lation using the criteria defined, the ratio of this patient's right adrenalvein to peripheral aldosterone levels are 2.06 in the unstimulatedphase and 2.69 in the stimulated phase. On thewhole, this gives reason-able evidence of success and on this basis the LI were calculated anddemonstrated clear right-sided dominance. Further, this subject wenton to have right adrenalectomy which cured his hypertension.

Use of ACTH-stimulation is controversial. Some sites have deter-mined that the risks of decreased LI outweigh the benefits of increasedSI [12,29] while other sites have advocated its use at sites with lowerprocedural success rates [30]. In the present study, there is no doubtof the utility of ACTH to improve the SI. In this sense, it has the abilityto “rescue” an otherwise uninterpretable procedure. However, thedown-side of ACTH is clear: in cases of PA showing apparent unilateraldisease based on unstimulated results, it has the undesirable effect ofdecreasing the LI and potentially yielding conflicting results betweenthe pre and post ACTH phases (whichwe observed in 14 of 44 potentialoccurrences). This phenomenon has been previously observed andappropriate concern expressed [11,12]. Armed with this knowledge,in practice, it is our opinion that the benefits of ACTH-stimulationoutweigh the costs because it rescues some procedures from beinguninterpretable. Specifically, in 117 cases (from sites C and E) forwhich collections were performed with and without stimulation,ACTH rendered 36 cases interpretable based on improved SI results ofthe post-ACTH phase: 21were lateralized (9 right, 12 left), 3 were inde-terminate and 12 bilateral. Obviously, because ACTH decreases the LI incases of unilateral PA by stimulating the unaffected adrenal gland[11,12], we can have clinical confidence in the 21 lateralized casesonly. Accordingly, in cases where only the post-ACTH results meetcannulation criteria, and the associated interpretation is indeterminateor bilateral, it is incumbent upon the interpreter to consider whetherthere is some reasonable evidence of cannulation in the pre-ACTHphase (based on venous aldosterone/cortisol results) and whether theinterpretation should be based preferentially on those data. Failingthis, a repeat AVS procedure should be considered.

In addition to ACTH-stimulation, a number of measures have beenproposed to improve AVS success. These have included the use of:CT-angiogram prior to AVS to determine the thoracic level of theRAV's entry to the IVC [8], standard operating procedures, rapidintra-procedural immunoassay for cortisol to confirm cannulation[31–33] and specifically selected catheters with appropriate calibreand a side-hole proximal to the tip [8]. Another measure we haveundertaken is to provide a formal report to the radiologist. Similar toprevious observations [24], sites with the lowest average cases peroperator (Sites A, B and F) had the poorest success rates. This under-scores the need for hospitals to designate one or two radiologists tobecome their site’s sole operators. If a critical mass of cases cannot be

Table 5Example AVS results where catheterization criteria are not achieved but evidence oftechnical success is substantial.

SAC (pmol/L) Cortisol (nmol/L) SI A/C LI

RAV Pre 2143 375 1.47 5.7 5.7LAV Pre 1660 1654 6.49 1IVC Pre 1042 255 NA 4.1RAV Post 4953 733 1.51 6.8 7.6LAV Post 13995 15428 31.68 0.9IVC Post 1841 487 1 3.8

1404 P. Elliott, D.T. Holmes / Clinical Biochemistry 46 (2013) 1399–1404

attained, it is advisable for sites to refer patients to one performingmoreprocedures.

We were able to identify laboratory evidence of dynamic functiontesting in only 92 of 198 of our subjects, which suggests that patientsare being directed from screening to AVS without receiving a confirmeddiagnosis of PA. However, in some of these cases, the clinical presentationmay have been so severe that confirmatory testing was considered un-necessary. Alternatively, some physicians may be using fludrocortisonechallenge followed by SAC analysis, whichwould not have been apparentusing our data collection protocol. We also note a relatively high propor-tion of APA (53% and 46% of successful cannulations pre and post-ACTHrespectively) compared with ~30–35% reported elsewhere [3,4]. SincePA tends to bemore severe in APA [3], we suspect a bias towardsworkupof more severe cases in BC.

The weakness of this study is that there is no clinical informationbeyond the patient demographics and there are no data on routinebiochemical investigations. Finally, there are no corroborative adrenalCT information, nomedication history for screening data, no anatomicalpathology data and no post-surgical clinical and biochemical follow-up.However, our purposewas to investigate the success-rate of AVS proce-dures with an aim to provide clinicians and radiologists an objectivelook at our performance, for improved patient care and resourceutilization.

Conclusions

Technical success rates of AVS procedures in BC are lower thanreported elsewhere but similar to observations of the German ConnSyndrome Registry. The site with a single dedicated radiological oper-ator had the greatest success. ACTH-stimulation increases the SI at thecost of significant decreases in the LI. However, with careful interpre-tation, ACTH-stimulation is useful and unequivocally rescues somecases from yielding no clinical data. We recommend AVS collectionbefore and after ACTH with greater diagnostic credence lent tounstimulated results. Special caution is warranted for cases whereonly ACTH-stimulated results show cannulation-success and calcula-tions suggest BAH. In this circumstance, consideration should begiven to repeat collection.

Acknowledgements

We are indebted to Ms. Grace Van der Gugten for the developmentand maintenance of our LC–MS/MS aldosterone assay. We wouldalso like to acknowledge Ms. Norine Freedman, Ms. Lynn Coleman,Ms. Karen Green, and Mr. William Prest for performing aldosteroneanalyses. Finally, we are indebted to Drs. Ted Wilkins, Jake Onrot,and Frances Rosenberg for establishing the Conn's Syndrome diagnosticprotocol in BC.

References

[1] Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global burdenof hypertension: analysis of worldwide data. Lancet 2005;365:217–23.

[2] Rossi GP, Bernini G, Caliumi C, Desideri G, Fabris B, Ferri C, et al. A prospectivestudy of the prevalence of primary aldosteronism in 1,125 hypertensive patients.J Am Coll Cardiol 2006;48:2293–300.

[3] Young WF. Primary aldosteronism: renaissance of a syndrome. Clin Endocrinol(Oxf) 2007;66:607–18.

[4] Rossi GP. A comprehensive review of the clinical aspects of primary aldosteron-ism. Nat Rev Endocrinol 2011;7:485–95.

[5] 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: anendocrine society clinical practice guideline. J Clin EndocrinolMetab 2008;93:3266–81.

[6] Montori VM, Young WF. Use of plasma aldosterone concentration-to-plasmarenin activity ratio as a screening test for primary aldosteronism. A systematic re-view of the literature. Endocrinol Metab Clin North Am 2002;31:619–32.

[7] Mulatero P, Milan A, Fallo F, Regolisti G, Pizzolo F, Fardella C, et al. Comparison ofconfirmatory tests for the diagnosis of primary aldosteronism. J Clin EndocrinolMetab 2006;91:2618–23.

[8] Daunt N. Adrenal vein sampling: how to make it quick, easy, and successful.Radiographics 2005;25(Suppl. 1):S143–58.

[9] Solar M, Ceral J, Krajina A, Ballon M, Malirova E, Brodak M, et al. Adrenal venoussampling: where is the aldosterone disappearing to? Cardiovasc Intervent Radiol2010;33:760–5.

[10] Kempers MJE, Lenders JWM, Van Outheusden L, der Wilt GJ Van, Schultze Kool LJ,Hermus ARMM, et al. Systematic review: diagnostic procedures to differentiateunilateral from bilateral adrenal abnormality in primary aldosteronism. AnnIntern Med 2009;151:329–37.

[11] Rossi GP, Pitter G, Bernante P, Motta R, Feltrin G, Miotto D. Adrenal vein samplingfor primary aldosteronism: the assessment of selectivity and lateralization ofaldosterone excess baseline and after adrenocorticotropic hormone (ACTH)stimulation. J Hypertens 2008;26:989–97.

[12] Seccia TM, Miotto D, De Toni R, Pitter G, Mantero F, Pessina AC, et al. Adrenocor-ticotropic hormone stimulation during adrenal vein sampling for identifyingsurgically curable subtypes of primary aldosteronism: comparison of 3 differentprotocols. Hypertension 2009;53:761–6.

[13] Van Der Gugten JG, Dubland J, Liu H-F, Wang A, Joseph C, Holmes DT. Determinationof serum aldosterone by liquid chromatography and tandem mass spectrometry:a liquid–liquid extraction method for the ABSCIEX API-5000 mass spectrometrysystem. J Clin Pathol 2012;65:457–62.

[14] Poulsen K, Jorgensen J. An easy radioimmunological microassay of renin activity,concentration and substrate in human and animal plasma and tissues based onangiotensin I trapping by antibody. J Clin Endocrinol Metab 1974;39:816–25.

[15] Dorrian CA, Toole BJ, Alvarez-Madrazo S, Kelly A, Connell JMC, Wallace AM. Ascreening procedure for primary aldosteronism based on the Diasorin Liaisonautomated chemiluminescent immunoassay for direct renin. Ann Clin Biochem2010;47:195–9.

[16] Rossi GP, Barisa M, Allolio B, Auchus RJ, Amar L, Cohen D, et al. The Adrenal VeinSampling International Study (AVIS) for identifying the major subtypes of primaryaldosteronism. J Clin Endocrinol Metab 2012;97:1606–14.

[17] Schirpenbach C, Seiler L, Maser-Gluth C, Beuschlein F, Reincke M, Bidlingmaier M.Automated chemiluminescence-immunoassay for aldosterone during dynamictesting: comparison to radioimmunoassays with and without extraction steps.Clin Chem 2006;52:1749–55.

[18] Pizzolo F, Corgnati A, Guarini P, Pavan C, Bassi A, Corrocher R, et al. Plasma aldoste-rone assays: comparison between chemiluminescence-based and RIA methods.Clin Chem 2006;52:1431–2.

[19] Mosso L, Carvajal C, González A, Barraza A, Avila F, Montero J, et al. Primaryaldosteronism and hypertensive disease. Hypertension 2003;42:161–5.

[20] Kim KI, Van deWiel MA. Effects of dependence in high-dimensional multiple testingproblems. BMC Bioinformatics 2008;9:114.

[21] Nwariaku FE, Miller BS, Auchus R, Holt S, Watumull L, Dolmatch B, et al. Primaryhyperaldosteronism: effect of adrenal vein sampling on surgical outcome. ArchSurg 2006;141:497–502.

[22] Young WF, Stanson AW, Thompson GB, Grant CS, Farley DR, Van Heerden JA. Rolefor adrenal venous sampling in primary aldosteronism. Surgery 2004;136:1227–35.

[23] Doppman JL, Gill Jr JR. Hyperaldosteronism: sampling the adrenal veins. Radiology1996;198:309–12.

[24] Vonend O, Ockenfels N, Gao X, Allolio B, Lang K, Mai K, et al. Adrenal venoussampling: evaluation of the German Conn's registry. Hypertension 2001;57:990–5.

[25] Toniato A, Bernante P, Rossi GP, Pelizzo MR. The role of adrenal venous samplingin the surgical management of primary aldosteronism. World J Surg 2006;30:624–7.

[26] Rossi GP, Sacchetto A, Chiesura-Corona M, De Toni R, Gallina M, Feltrin GP, et al.Identification of the etiology of primary aldosteronism with adrenal veinsampling in patientswith equivocal computed tomography andmagnetic resonancefindings: results in 104 consecutive cases. J Clin Endocrinol Metab 2001;86:1083–90.

[27] Roberts RF, Roberts WL. Performance characteristics of five automated serumcortisol immunoassays. Clin Biochem 2004;37:3–489.

[28] Mulatero P, Bertello C, Sukor N, Gordon R, Rossato D, Daunt N, et al. Impact of differ-ent diagnostic criteria during adrenal vein sampling on reproducibility of subtypediagnosis in patients with primary aldosteronism. Hypertension 2010;55:3–667.

[29] Rossi GP, Ganzaroli C, Miotto D, De Toni R, Palumbo G, Feltrin GP, et al. Dynamictesting with high-dose adrenocorticotrophic hormone does not improve lateraliza-tion of aldosterone oversecretion in primary aldosteronism patients. J Hypertens2006;24:371–9.

[30] Monticone S, Satoh F, Giacchetti G, Viola A, Morimoto R, Kudo M, et al. Effect ofadrenocorticotropic hormone stimulation during adrenal vein sampling in prima-ry aldosteronism. Hypertension 2012;59:840–6.

[31] Auchus RJ, Michaelis C, Wians Jr FH, Dolmatch BL, Josephs SC, Trimmer CK, et al.Rapid cortisol assays improve the success rate of adrenal vein sampling for prima-ry aldosteronism. Ann Surg 2009;249:318–21.

[32] Betz MJ, Degenhart C, Fischer E, Pallauf A, Brand V, Linsenmaier U, et al. Adrenalvein sampling using rapid cortisol assays in primary aldosteronism is useful incenters with low success rates. Eur J Endocrinol 2011;165:301–6.

[33] Stowasser M. Improving the success and reliability of adrenal venous sampling:focus on intraprocedural cortisol measurement. Clin Chem 2012;58:1275–7.