Biochemical Diagnosis of Pheochromocytoma- Which Test is Best.full

8/2/2019 Biochemical Diagnosis of Pheochromocytoma- Which Test is Best.full

1/8


2/8


3/8

was performed prospectively in 819 ofthe 858 patients included in the finalanalyses. In the remaining patients, allwith pheochromocytoma, measure-ments of plasma free metanephrineswere performed after removal of tu-

mors. Twenty-four hour urine collec-tions were obtainedfrom 790 of the858patients included in the final analyses.Urine samples were usually subjectedto assays of 2 to 3 different analytes.

Plasma was analyzed, using high-performance liquid chromatography(HPLC), for concentrations of catechol-amines, norepinephrine and epineph-rine, and free metanephrines, normeta-nephrine, and metanephrine.1 7 - 1 9

Metanephrines in urine were ana-lyzed as fractionated normetaneph-rine and metanephrine by HPLC or as

total metanephrines by spectrophotom-etry.10,20 Urinary VMA was measuredspectrophotometrically and urinary cat-echolamines by HPLC.21,22 Upper ref-erence limits for biochemical tests areprovided as established by the princi-pal laboratories responsible for eachtest(TABLE 2).11

All 858 assays of plasma free meta-nephrines and 91% of the 855 assaysof plasma catecholamines were per-formed at the National Institutes ofHealth. Seventy-five percent of the 557

assays of urinary fractionated meta-nephrineswere performed by Quest Di-agnostics (Collegeville, Pa). Seventy-two percent of the 710 assays of urinarycatecholamines, 74% of the 297 as-saysof urinary total metanephrines, and79% of the 616 assays of urinary VMAwereperformed by MayoMedical Labo-ratories (Rochester, Minn). The re-maining assays were performed at dif-ferent laboratories depending on thecenter where patients were evaluated.

Data Analysis

For biochemical tests involving pairs ofmeasurements (eg, normetanephrineand metanephrine or norepinephrineand epinephrine in plasma or urine),a false-negative result in a patientwith pheochromocytoma or a true-negative result in a patient withoutpheochromocytoma was defined as a

value for each measurement lower thanthe upper reference limit. A true-positive result for pairs of measure-ments in a patient with pheochromo-cytoma or a false-positive result in apatient without pheochromocytoma

was defined as a value for either or bothmeasurements equal to or higher thanthe appropriate upper reference limit.

Sensitivity was calculated from thepercentageof true-positiveover the totalof true-positive plus false-negative testresults in patients with pheochromo-cytoma. Specificity was calculatedfromthepercentage of true-negativeover thetotal of true-negative plus false-positive test results in patients with-out pheochromocytoma. Differencesinsensitivity and specificity were exam-ined using the McNemar test and are

illustrated using 95% confidence inter-vals (CIs).

For each analyte, a receiver operat-ing characteristic (ROC) curve wascon-structed from the relationship be-tween true-positive and false-positiveresults (ie, sensitivity vs 1specific-ity) at different upper reference limits

for each analyte.23 As summary mea-sures of the diagnostic utility of eachtest, independent of upper referencelimits, areas under ROC curves werecalculated and differences among testsexamined according to the method by

Hanley and McNeil.

24

Negative predictive values were cal-culated from the percentage of true-negative over the total of true-negativeplus false-negative test results. Positivepredictive values were calculated fromthe percentage of true-positive over thetotal of true-positive plus false-positiveresults. Positive and negative predic-tive values of each test were calculatedat different prevalences of pheochromo-cytoma to establish posttest probabili-ties of pheochromocytoma at differentpretest probabilities of the tumor.

RESULTS

Biochemical Test Results

Relative to patients in whom pheochro-mocytoma wasexcluded, medianplasmaconcentrations of free metanephrineswere increased by 7-foldin patients withhereditary pheochromocytoma and 21-

Table 2. Plasma Concentrations of Catecholamines and Metanephrines and Urinary Outputsof Catecholamines, Metanephrines, and Vanillylmandelic Acid

UpperReference

Limit

PheochromocytomaConfirmed*

PheochromocytomaExcluded*

Hereditary Sporadic Hereditary Sporadic

Plasma, nmol/LFree metanephrines

Normetanephrine 0.6 2.2 (0.3-69) 6.3 (0.4-173) 0.3 (0.1-1.6) 0.3 (0.1-2.2)

Metanephrine 0.3 0.2 (0-15) 0.6 (0-383) 0.1 (0-0.5) 0.2 (0-1.2)

CatecholaminesNorepinephrine 2.9 4.8 (1-80) 10.6 (0.7-1360) 1.4 (0.4-8.1) 2.1 (0.3-29)

Epinephrine 0.5 0.2 (0-4.8) 0.4 (0-1111) 0.1 (0-1.3) 0.1 (0-18)

Urine, mol/dFractionated

metanephrinesNormetanephrine

Women 1.75.2 (1.8-33) 17.7 (0.9-191) 1.2 (0.1-8.2) 2.1 (0.4-10)

Men 3.0

MetanephrineWomen 0.7

0.5 (0.1-28) 1.1 (0.1-466) 0.4 (0.1-2.8) 0.5 (0.1-22)Men 1.2

CatecholaminesNorepinephrine 0.5 1 (0.1- 5.4) 1.7 (0.1- 31) 0.2 (0-1) 0.3 (0.1-5.6)

Epinephrine 0.1 0 (0-2.1) 0.1 (0-7) 0 (0-0.1) 0 (0-0.4)

Total metanephrines 6 8 ( 2-260) 21 (3-234) 3 (1-12) 3 (1-16)

Vanillylmandelic acid 40 37 (11-330) 83 (13-777) 17 (3-56) 23 (2-81)

*Values are expressed as median (range).Measured by high-performance liquid chromatography after acid deconjugation.Measured together as a single analyte by spectrophotometry after acid deconjugation.

DIAGNOSIS OF PHEOCHROMOCYTOMA

2002 American Medical Association. All rights reserved. (Reprinted) JAMA, March 20, 2002Vol 287, No. 11 1429


4/8

fold in patients with sporadic pheochro-mocytoma (Table 2). These increaseswere consistently larger than those ofplasma norepinephrine (3-foldand5-foldincreases), urinary norepinephrine (5-fold and 6-fold increases), urinary frac-tionated normetanephrine (4-fold and8-fold increases), urinary total meta-nephrines (3-fold and 7-fold in-creases), and urinary VMA (2-fold and4-fold increases). Increases in all ana-lytes were larger in patients with spo-radic rather than hereditary pheochro-

mocytoma.

Test Sensitivity

The sensitivities of diagnostic tests fordetection of hereditary or sporadicpheochromocytoma ranged from a lowof 46% (95% CI, 34%-59%) for use ofurinary VMA in the detection of he-reditary pheochromocytoma to a highof 99% (95% CI, 96%-100%) for use ofplasma free metanephrines in the de-tection of sporadic pheochromocy-toma (TABLE 3).

Amongall patients withpheochromo-cytoma, sensitivities were thehighest formeasurements of plasma free metaneph-rines at 99% (95% CI, 96%-100%), fol-lowed closely by urinary fractionatedmetanephrines at 97% (95% CI,92%-99%).Sensitivitiesof both theabovetests considerably (P.001) exceededthose for urinary catecholaminesat 86%

(95% CI, 80%-91%), plasma catechol-amines at 84% (95% CI, 78%-89%), uri-nary total metanephrines at 77% (95%CI, 68%-85%), and urinary VMA at 64%(95% CI, 55%-71%).

Theabove variations in sensitivities ofdiagnostic tests showed similar pat-terns inpatients with hereditary andspo-radic pheochromocytoma (Table 3).Plasma free metanephrines and uri-nary fractionated metanephrines of-fered the highest sensitivities. Plasmaand urinary catecholamines had inter-

mediate sensitivities. Urinarytotal meta-nephrines andVMAconsistentlyshowedthe lowest sensitivities in both groupsof patients. The sensitivities of all testswere higher for detection of sporadicpheochromocytoma than for detectionof hereditary pheochromocytoma.

Test Specificity

Specificities of biochemical testsrangedwidely from 45% (95% CI, 36%-51%)for urinary fractionated metaneph-rines in patients tested for sporadic

pheochromocytoma to 99% (95% CI,98%-100%) for urinary VMA in pa-tients tested for hereditary pheochro-mocytoma (Table 3).

Among all patients tested for pheo-chromocytoma, the highest specifici-ties were 95% (95% CI, 93%-97%) fortests of urinaryVMA and 93% (95% CI,89%-97%)fortests of urinary total meta-

nephrines. Specificities were interme-diate for tests of plasma free metaneph-rines at 89% (95% CI, 87%-92%),urinarycatecholamines at 88%(95% CI,85%-91%), and plasma catechol-amines at 81% (95% CI,78%-84%), and

lower than those of all other tests(P.001) for urinary fractionatedmeta-nephrines at 69% (95% CI, 64%-72%).

In contrast to sensitivities, specifici-ties of all tests were higher in patientstested for hereditary pheochromocy-toma than for sporadic pheochromo-cytoma (Table 3). In both groups, uri-nary VMA and total metanephrinesoffered the highest specificities and uri-nary fractionated metanephrines thelowest specificities.

False-Negative Plasma

Free MetanephrinesOnly 2 of the 76 patients with heredi-tarypheochromocytoma and1 ofthe138patients with sporadic pheochromocy-toma had normal levels of plasma freemetanephrines. Both hereditary caseswere in patients who were normoten-sive andasymptomaticandhadno otherbiochemical evidence ofthetumor. Bothwere patients with von Hippel-Lindausyndromeandhad singleadrenal tumorsoflessthan1cmindiameter,whichwereidentified and removed coincidentally

during surgery for renal carcinoma.The single false-negative result for

plasma free metanephrines in patientswith sporadic pheochromocytoma in-volved a patient tested for possible tu-mor recurrence 13 years after theremoval of a large extra-adrenal pheo-chromocytoma. Computed tomogra-phy and all biochemical tests yieldednegative results. The patient was subse-quently diagnosedwithmetastaticpheo-chromocytoma 312 years later. Sincethere was no evidence for a hereditary

basis for thepatients disease, it was pre-sumed that the malignancy developedsecondary to remaining diseasethatwentundetected for more than 16 years afterthe original tumor was removed. Thus,despite the considerable time betweenbiochemical testing and final diagnosis,all tests performed were designated asproviding false-negative results.

Table 3. Sensitivities and Specificities of Biochemical Tests for Diagnosis of Hereditary andSporadic Pheochromocytoma*

Sensitivity, % Specificity, %

Heredi tary Sporadi c Hereditary Sporadic

PlasmaFree metanephrines 97 (74/76) 99 (137/138) 96 (326/339) 82 (249/305)

Catecholamines 69 (52/ 75) 92 ( 126/137) 89 (303/339) 72 (220/ 304)Urine

Fractionated metanephrines 96 (26/27) 97 (76/78) 82 (237/288) 45 (73/164)

Catecholamines 79 (54/68) 91 (97/107) 96 (312/324) 75 (159/211)

Total metanephrines 60 (27/45) 88 (61/69) 97 (91/94) 89 (79/89)

Vanillylmandelic acid 46 (30/65) 77 (66/86) 99 (310/312) 86 (132/153)

*The reference limits used to calculate sensitivity and specificity are presented in Table 2.For free plasma metanephrines or urinary fractionated metanephrines, sensitivity was calculated from patients with

pheochromocytoma and false-negative test results for both normetanephrine and metanephrine. For plasma andurine catecholamines, sensitivity was calculated from patients with both false-negative test results for nonrepineph-rine and epinephrine. Numbers in parentheses indicate true positive over true positive plus false-negative.

For free plasma metanephrines or urinary fractionated metanephrines, specificity was calculated from patients with-out pheochromocytoma and with false-positive test results for either normetanephrine or metanephrine. For plasmaand urine catecholamines, specificity was calculated from patients without pheochromocytoma and with false-positive test results for either nonrepinephrine or epinephrine. Numbers in parentheses indicate true negative overtrue negative plus false-positive.


1430 JAMA, March 20, 2002Vol 287, No. 11 (Reprinted) 2002 American Medical Association. All rights reserved.


5/8

ROC Curves

Integrated comparisonof sensitivity andspecificity using ROCcurvesshowedthatthediagnostic power ofplasmafreemeta-nephrineswassuperior tothat ofallothertests (FIGURE 1). The areas under the

ROC curves for plasma catecholamines(0.927),urinary catecholamines (0.931),urinary total metanephrines(0.919), andurinary VMA (0.896) were all signifi-cantly lower (P.001) than the area forplasma free metanephrines (0.985). Al-though closer, the area under the ROCcurvefor urinary fractionatedmetaneph-rines (0.960) was also lower (P=.01)than that for plasma free metaneph-rines (0.985).

Areas under the ROC curves wereonly marginally improved when testsof urinary fractionated metanephrines

were combined with those for urinarycatecholamines (0.965) or plasma cat-echolamines (0.969) or when tests ofurinary total metanephrines and cat-echolamines were combined (0.949)(Figure 1). Thus, combining tests fordifferent analytes did not improve di-agnostic efficacy beyond that of a singletest of plasma free metanephrines.

True-positive rates (ie, test sensitiv-ity) at higher upper reference limitswhen false-positive rates were zero (ie,whentestspecificityequaled100%) were

higher for plasma free metanephrinesthan for other tests (Figure 1). None ofthe 644 patients without pheochromo-cytoma had a plasma concentration ofnormetanephrine above 2.19 nmol/L orof metanephrine above 1.20 nmol/L,whereas79%of patientswith pheochro-mocytomahad plasma concentrationsofnormetanephrine or metanephrineabove these levels (TABLE 4).

Positive and Negative

Predictive Values

Negative predictive values of tests ofplasma and urinary metanephrines atdifferent prevalences of pheochromo-cytoma showed that negative test re-sults for plasma free and urinary frac-tionated metanephrines provided thehighest probabilities for excludingpheochromocytoma at all pretest preva-lences of the tumor (FIGURE 2). How-

ever, the posttest probability of a pheo-chromocytoma from a positive testresult for plasma free metanephrines,although similar to that for urinary totalmetanephrines, was consistently higherthan that from a positive test result for

urinary fractionated metanephrines atall pretest prevalences of the tumor.

COMMENT

The present examination of biochemi-cal tests used in the diagnosis ofpheochromocytoma provides severaladvances overpreviousstudies.First,thisstudycomprehensively compared mea-surements of plasma free metaneph-rines with all other commonly avail-able biochemical tests used to diagnoseexcess catecholamine production. Sec-ond, these comparisons were made in

large populations of patients with andwithoutpheochromocytoma, whoweretested for the tumor because of clini-cally appropriate predisposing condi-tions or suspicioussymptoms andsigns.Finally, standardcriteriathatwereinde-pendent of the biochemical tests beingcompared were used to assign patientsintogroups with andwithout thetumor.

Sensitivity, Specificity,

and ROC Curves

The present study confirms the find-

ings of several other reports thatmeasurements of plasma free meta-nephrinesor urinary fractionated meta-nephrines offer higher sensitivity for di-agnosis of pheochromocytoma thanmeasurements of plasma or urinary cat-echolamines or of urinary total meta-nephrines or VMA.8,9,11,14,16 Our com-parisons further establish that amongall tests, including urinary fraction-ated metanephrines, measurements ofplasma free metanephrines provide thebest test for excluding or confirming

pheochromocytoma.Since measurements of urinary frac-tionated metanephrines andplasmafreemetanephrines offer similarly high sen-sitivity, a negative result for either testis equally effective for excluding pheo-chromocytoma. However, because uri-nary fractionated metanephrines havelow specificity, tests of plasma free

Figure 1. Receiver Operating CharacteristicCurves

True-PositiveRate

(TestSensitivity)

1.0

0.6

0.8

0.4

0.2

0

0 0.2 0.6

Plasma Free Metanephrines

Plasma Catecholamines

Urinary Catecholamines

Urinary Vanillylmandelic Acid

1.00.80.4

1.0 0.8 0.4 00.20.6

True-Negative Rate (Specificity)

Plasma Free Metanephrines vs Urinary

Catecholamines and Vanillylmandelic Acid

A

Plasma Free Metanephrines vs Urinary

Fractionated Metanephrines and Total

Metanephrines

B


vs Test Combination

C

True-Positive

Rate(TestSensitivity)

1.0

0.6

0.8

0.4

0.2

0

0 0.2 0.6


Urinary Fractionated Metanephrines

Urinary Total Metanephrines

1.00.80.4

1.0 0.8 0.4 00.20.6

True-PositiveRate

(TestSensitivity)

1.0

0.6

0.8

0.4

0.2

0

0 0.2 0.6


Urinary Fractionated Metanephrinesand Plasma Catecholamines

Urinary Total Metanephrinesand Urinary Catecholamines

1.00.80.4

1.0 0.8 0.4 00.20.6

False-Positive Rate (1 Specificity)

Urinary Fractionated Metanephrinesand Urinary Catecholamines

Relationships between rates of true-positive test re-sults (ie, sensitivity) andratesof false-positive test re-sults (1 specificity) were calculated at different up-perreferencelimitsfor each ofthe tests.At higherupperreference limits, rates of true-positive test results de-crease(ie, sensitivitydecreases),whereas rates of false-positive test results increase (ie, specificity increases).




6/8

metanephrines exclude pheochromo-cytoma in many more patients with-

out the tumor than do tests of urinaryfractionated metanephrines.The above considerations illustrate the

importance of ROC curves for compar-ing different tests. At equivalent levelsof sensitivity, the specificity of plasmafree metanephrines is higher than thatof other tests. At equivalent levels ofspecificity, the sensitivity of plasma free

metanephrines is also higher than thatof other tests, including urinary frac-tionated metanephrines.

Multiple Biochemical Tests

To minimize the risk of missing a pa-tient with pheochromocytoma, clini-cians often use multiple biochemical

tests during the initialdiagnostic workupof patients with suspected tumors. Al-though this may increase sensitivity, itis at the cost of decreased specificity.Thus, tests involving pairs of measure-ments, such as fractionated catechol-amines or metanephrines, have lowerspecificity and higher sensitivity thantests involving single measurements,such as urinary VMA or total metaneph-rines (Table 3).8 As shown by ROCcurves, the diagnostic utility of tests ofplasma free metanephrines remains su-

perior to that of other tests even whenthe latter are used in combination.The above considerations lead us to

recommend against use of multiplebio-chemical tests to exclude pheochro-mocytoma in favor of a single test ofplasma free metanephrines. In pa-tients with negative test results forplasma free metanephrines, indiscrimi-nate use of extra tests is unlikely to im-prove diagnostic efficacy. If multiplebiochemical tests have been run, thede-cision to exclude pheochromocytomashould be basedon whether plasma freemetanephrines show a negative test re-sult, regardless of whether other test re-sults are positive or negative.

Differences in Test

Performance Explained

Whydo plasmafreemetanephrinespro-vide the best test to diagnose pheochro-

mocytoma? First, plasma free meta-nephrines are produced continuouslyby metabolismof catecholamineswithinpheochromocytoma tumor cells.25-27

This contrasts with episodic secretionof catecholamines. Second, sympa-

thoadrenal excitation causes largeincreases in catecholamine release,whereas plasma free metanephrines re-mainrelatively unaffected.14,25,27,28 Third,VMA and the total and fractionatedmetanephrines measured in urine aredifferent metabolites from the freemeta-nephrines measured in plasma, and areproduced in different parts of the bodyby metabolic processes not directly re-lated to the tumor itself.28-30 Urinarytotal and fractionated metanephrinesaremeasured after a deconjugation stepand largely reflect levels of conjugated

metanephrines that are produced out-side of tumor tissue. Similarly, VMA isproduced mainly in the liver.

Sporadic vs Hereditary

Pheochromocytoma

The lower sensitivity and higher speci-ficity of biochemical tests for hereditarycompared with sporadic pheochromo-cytoma reflect differentreasons for test-ing in the 2 groups.31 Routine screeningfor pheochromocytoma in patientswitha hereditary predisposition to the tu-

mor often leadsto detection of small tu-mors that release catecholamines inamounts that are insufficient to pro-duce signs or symptomsof the tumor. Incontrast,sporadicpheochromocytomaistypically suspected because of signs andsymptomsof catecholamine excess, pro-duced by larger more easilydetectedtu-mors thanfound by routinescreening inhereditary pheochromocytoma. More-over, patients tested for sporadic pheo-chromocytoma who do not have the tu-mor are often symptomatic of some

condition associated with sympathoad-renal activation,leadingto relatively highnumbers of false-positive results.

The consistently lower specificities ofbiochemical tests in patients tested forsporadic rather than for hereditarypheochromocytomamay alsoreflect re-ferral of patients in the former groupwith previously determined positive

Table 4. Plasma Concentrations, Specificity, and Sensitivity of Plasma Free Metanephrines atEstablished Upper Reference Limits Compared With Reference Limits Adjusted to Provide aZero False-Positive Rate

Upper Reference Limits Plasma Level, nmol/L Test Specificity, % Test Sensitivity, %

EstablishedNormetanephrine 0.61

89 99Metanephrine 0.31

Adjusted*Normetanephrine 2.19

100 79Metanephrine 1.20

*Plasma concentrations and test sensitivity adjusted to provide a zero false-positive rate were established from re-ceiver operating characteristic curves in Figure 1.

Figure 2. Relationships Between Pretest andPosttest Probability

PosttestProbability

1.0

0.6

0.8

0.4

0.2

0

0 0.2 0.6




1.00.80.4

Pretest Probability (Prevalence)

Plasma Free MetanephrinesPositive Predictive Values

Negative Predictive Values



Positive

Predict

iveValue

s

1Ne

gativ

ePredict

iveValue

s

A positive or a negative result for tests of plasma orurinary metanephrines changes the respective prob-abilities of having or not having pheochromocytomain relationshipto differentpretestprobabilities (preva-lences) of the tumor. The dotted line represents therelationship expected between pretest and posttestprobability if a test hadno diagnostic value.Relation-ships illustrated by the filled symbols show the prob-abilities of havinga pheochromocytomabased on posi-tive(abnormal)test results,whereasthe open symbolsshow the probabilities of not having a pheochromo-cytoma based on negative (normal) test results.




7/8

biochemical tests. Thus, specificities ofbiochemical tests for detection of spo-radic pheochromocytoma in the pres-ent study are likelyto belower than inunselected populations tested by com-mercial laboratories, but should re-

flect those expected in populationstested at referral centers.Apart from patients at risk for heredi-

tary pheochromocytoma, patients withpreviously resected tumors are anotherat-risk group who should be tested pe-riodically for the tumor regardless ofsigns and symptoms. The importance ofthis group is illustratedby thesingle pa-tient who tested negative for pheochro-mocytoma by all tests 312 years beforemetastatic disease was finally diag-nosed and 16 years after removal of theprimary tumor.Thesensitivity of plasma

free metanephrines is not always suffi-cient for detection of microscopic re-current or metastatic disease or smalltumors (1 cm) in patientswith heredi-tary pheochromocytoma.

Pretest and Posttest Probabilities

Typically fewer than 1% of hyperten-sive patients tested for pheochromo-cytoma have the tumor. In some pa-tient groups, such as those withhypertension and an adrenal mass, thepretest probability of a pheochromo-

cytoma may be higher. The probabil-ity that a negative test result excludespheochromocytoma or that a positivetest result confirms the tumor de-pends in part on the pretest probabil-ity of the disease. These posttest prob-abilitiestherefore require calculation ofpositive and negative predictive val-ues at different pretest probabilities(prevalences) of the tumor.

As shown in Figure 2, a negative testresult for plasmafree metanephrines orurinaryfractionatedmetanephrines pro-

vides a high probability of excludingpheochromocytoma at all clinically rel-evant pretest probabilities of the tu-mor. In contrast, at the typically lowprevalences of pheochromocytoma, thelikelihood of the tumor after a singlepositive test remains low even for testswith up to 95% specificity. A routinepractice to further increase or de-

crease the likelihood of pheochromo-cytoma involves use of additional bio-chemical tests.32

In patients with positive results foran initial test of plasma free metaneph-rines,extratests canbe useful, butjudg-

ing the likelihood of a pheochromocy-toma should first take into accountresults of ROC curves. In particular, athigher upper reference limits, in whichtest specificity is 100%, plasma con-centrations of free normetanephrinehigher than 2.19 nmol/L or of free meta-nephrine higher than 1.20 nmol/L un-equivocally confirm a pheochromocy-toma in 79% of patients with the tumor(Table 4), a higher proportion than forother tests (Figure 1). The probabilityof pheochromocytoma in these pa-tients is so high that further biochemi-

cal tests to confirm the tumor may beunnecessary. In the remaining pa-tients, in whom increased levels are in-sufficient to unequivocally confirm a tu-mor, additional judiciously selectedfollow-up tests are appropriate, withad-ditional attention focused on possiblecauses of false-positive test results.32-35

False-Positive Test Results

Because of the low prevalence of pheo-chromocytoma in the patient groupsusually tested for the tumor, false-

positive results can be expected to out-number true-positive results for all bio-chemical tests, including plasma freemetanephrines. There are 3 potentialsourcesof false-positive testresults: diet,drugs, and stressors.

Caffeic acid, a catechol found in cof-fee(including decaffeinated coffee), andits derivative dihydrocaffeic acid are di-etary substances known to interferewithassays of plasma catecholamines.36

Moreover, both catechols are excellentsubstrates for catechol O-methyltrans-

ferase, the enzyme that converts cat-echolaminesto metanephrines, and thiseasily could affect plasmalevels of meta-nephrines. There are many other uni-dentified dietary constituentsthatcanin-fluence the results of HPLC assays. Thesimplest way to avoid these sources offalse-positive results is by drawingbloodsamples in patients who have fasted.

Acetaminophen is the only directsource of interference with assays ofplasmafreemetanephrines thatwe haveidentified to date.19 However, caffeineand nicotine both increase plasma lev-els of catecholamines and should also

be avoided. In our series, treatment withtricyclic antidepressants or phenoxy-benzamine (dibenzyline) were majorcauses of false-positive test results fornorepinephrineanditsmetabolites, pre-sumably due to presynaptic actions onsympathetic nerves. Phenoxybenza-mine, a nonselective -adrenoceptorblocker commonly used to treat pa-tients with pheochromocytoma, can beparticularly troublesome.

Although plasma levels of free meta-nephrines are less sensitive to changesin sympathoadrenalactivity than arelev-

els of the parent amines, these metabo-lites are neverthelessinfluenced by manyof the same stimuli and drugs that in-fluence plasma catecholamines.25-28 Up-right posture and emotional stress arewell-known to stimulate release of cat-echolamines from sympathetic nervesand the adrenal medulla. To minimizethe possibility of false-positive test re-sults, we collected blood samples forplasma free metanephrines under thesame conditions used for collection ofsamplesfor measurementsof plasma cat-

echolamines. Blood samples weredrawnwith patients in the supine position,through an in-dwelling intravenouscatheter, and after an overnight fast.

Study Limitations

The major strength of this studyallpatientswere examined becauseof clini-cal suspicion of pheochromocytomawas associated with the limitation thatexclusion of pheochromocytomarequired use of methods other than thebiochemicaltestsnormally used in clini-

cal practice. Although computedtomog-raphy and magnetic resonance imagingoffer high sensitivity for detecting adre-nal tumors, sensitivity decreases fordetectingextra-adrenaldisease.Wethere-fore followed up patients for an averageof2.5yearsto furtherexclude pheochro-mocytoma in patients with negativeimaging. Only 1 patient in whom pheo-




8/8

chromocytoma wasinitially excludedbyimaging studieswassubsequentlyfoundto have the disease at follow-up. Itremains possible, however, that otherpatients in whom pheochromocytomawasexcluded according to thecriteriaof

ourstudymayhavehadundiagnoseddis-ease. Even so andunless these numberswere large, it is unlikely that incorrectcategorization of these patients wouldmake asignificantdifferenceto theresultsand conclusions of the study.

A related potential limitation of thestudy was the need to omit from theanalyses 145 patients who did not meettheresearch criteriaforexclusion or con-firmation of pheochromocytoma. Sepa-rate analyses of howinclusion of thedatafrom these patients would affecttestper-formance revealed little influence on the

results and conclusions of the study.Another potential limitation of the

study involved the multicenter nature ofpatientrecruitmentandsubsequent mea-surements of urinary analytes by differ-entlaboratories compared with thesinglelaboratory used for plasma free meta-nephrines.However,separate analysisofdata derived from single laboratoriesre-vealed no obvious influences. Thus,rather than a limitation, the multi-center nature of the study is a strengthestablishing thatmany ofthefindings(eg,

low specificity of urinary fractionatedmetanephrines, low sensitivity of uri-nary VMA) were independent of thelaboratory where tests were run.

Conclusions

Plasmafreemetanephrines constitute thebest test for excluding or confirmingpheochromocytoma and should be thetest of first choice for diagnosis of thetu-mor. A negative test result virtually ex-cludes pheochromocytoma.In such pa-tients, representing more than 80% of

those tested, no immediate further testsfor the tumor are necessary. Further-more, in about 80% of patients withpheochromocytoma, the magnitude ofincrease in plasmafreemetanephrinesisso large that the tumor can be con-firmed with close to 100% probability.In these patients, the immediate task isto locate the tumor.

Author Contributions: Study concept and design:Lenders, Walther, Friberg, Keiser, Eisenhofer.Acquisition of data: Lenders, Pacak, Walther, Man-nelli, Friberg, Eisenhofer.Analysis and interpretation of data: Lenders, Line-han, Goldstein, Eisenhofer.Drafting of the manuscript: Lenders, Linehan, Gold-stein, Eisenhofer.Critical revision of the manuscript for important in-

tellectual content: Lenders, Pacak, Walther, Man-nelli, Friberg, Keiser, Goldstein, Eisenhofer.Statistical expertise: Goldstein.Administrative, technical, or material support: Wal-ther, Friberg, Keiser, Goldstein, Eisenhofer.Study supervision: Lenders, Walther, Linehan, Gold-stein.Acknowledgment: We are grateful to CourtneyHolmes, BS, Patricia Sullivan, BS, and Jacques J. Wil-lemsen, BS, for expert technical assistance, to DonnaPeterson, RN, BS, and Svisse Haegelind, BS, for datamanagement, to William M. Manger, MD, PhD, forhelpful advice, toRaymondR. Townsend, MD,for as-sistance with patients, and to the many other clini-cians andtheirpatientswho participatedin thestudy.

REFERENCES

1. Bravo EL. Evolving concepts in the pathophysiol-

ogy, diagnosis, and treatment of pheochromocy-toma. Endocr Rev. 1994;15:356-368.2. Manger WM, Gifford RW Jr. Pheochromocy-toma: current diagnosis and management. Cleve ClinJ Med. 1993;60:365-378.3. Mannelli M. Diagnosticproblems in pheochromo-cytoma. J Endocrinol Invest. 1989;12:739-757.4. Bravo EL,Tarazi RC, GiffordRW, Stewart BH.Cir-culating and urinary catecholamines in pheochromo-cytoma. N Engl J Med. 1979;301:682-686.5. Duncan MW, Compton P, Lazarus L, Smythe GA.M e a s u r e m e n t o f n o r e p i n e p h r i n e a n d 3 , 4 -dihydroxyphenylglycol in urineand plasma for the di-agnosis of pheochromocytoma. N Engl J Med. 1988;319:136-142.6. Young MJ,Dmuchowski C, Wallis JW, BarnasGP,Shapiro B. Biochemical tests for pheochromocytoma:strategiesin hypertensive patients.J Gen Intern Med.1989;4:273-276.

7. Peaston RT, Lai LC. Biochemical detectionof phae-chromocytoma. J Clin Pathol. 1993;46:734-737.8. Gerlo EA, Sevens C. Urinary and plasma catechol-amines and urinary catecholamine metabolites in pheo-chromocytoma. Clin Chem. 1994;40:250-256.9. Lenders JW,KeiserHR, GoldsteinDS, et al.Plasmametanephrines in the diagnosis of pheochromocy-toma. Ann Intern Med. 1995;123:101-109.10. HeronE, Chatellier G, Billaud E, Foos E, PlouinPF.The urinary metanephrine-to-creatinine ratio for thediagnosis of pheochromocytoma. Ann Intern Med.1996;125:300-303.11. Eisenhofer G, Lenders JW, Linehan WM, Wal-ther MM, Goldstein DS, Keiser HR. Plasma normeta-nephrine and metanephrine for detecting pheochro-mocytoma in von Hippel-Lindau disease and multipleendocrine neoplasia type 2. N EnglJ Med. 1999;340:1872-1879.12. MannelliM, Ianni L, Cilotti A, Conti A. Pheochro-

mocytoma in Italy:a multicentric retrospective study.Eur J Endocrinol. 1999;141:619-624.13. Witteles RM, Kaplan EL, Roizen MF. Sensitivityof diagnostic and localization tests for pheochromo-cytoma in clinical practice. Arch Intern Med. 2000;160:2521-2524.14. Raber W, Raffesberg W, Bischof M, et al. Diag-nostic efficacy of unconjugated plasma metaneph-rines for the detection of pheochromocytoma. ArchIntern Med. 2000;160:2957-2963.15. Hernandez FC,Sanchez M, AlvarezA, et al.A five-year report on experience in the detection of pheo-

chromocytoma. Clin Biochem. 2001;33:649-655.16. Gardet V, Gatta B, Simonnet G, et al.Lessons froman unpleasant surprise. J Hypertens. 2001;19:1029-1035.17. Eisenhofer G, Goldstein DS, Stull R, et al. Simul-taneousliquid-chromatographic determination of 3,4-dihydroxyphenylglycol, catecholamines, and 3,4-dihydroxyphenylalanine in plasma,and their responsesto inhibition of monoamine oxidase.Clin Chem.1986;

32:2030-2033.18. van der Hoorn FAJ, Boomsma F, Man int VeldAJ, Schalekamp MADH. Determination of catechol-amines in human plasma by high-performance liquidchromatography. J Chromatogr. 1989;487:17-28.19. Lenders JWM, Eisenhofer G, Armando I, KeiserHR, Goldstein DS, Kopin IJ. Determination of plasmametanephrines by liquid chromatography with elec-trochemical detection. Clin Chem. 1993;39:97-103.20. Pisano JJ. A simple analysis of normetanephrineand metanephrine in urine. Clin Chim Acta. 1960;5:406-414.21. PisanoJJ, Crout R, Abraham D. Determination of3-methoxy-4-hydroxymandelicacid in urine. ClinChimActa. 1962;7:285-289.22. Moyer TP,Jiang NS,Tyce GM,ShepsSG. Analy-sis for urinary catecholamines by liquid chromatog-raphy with amperometric detection. Clin Chem.1979;25:256-263.

23. Beck JR, Shultz EK. The use of relative operatingcharacteristic (ROC)curvesin testperformanceevalu-ation. Arch Pathol Lab Med. 1986;110:13-20.24. Hanley JA, McNeil BJ. A method of comparing theareasunder receiver operating characteristiccurves de-rivedfrom thesamecases. Radiology. 1983;148:839-843.25. Eisenhofer G, Friberg P, Pacak K, et al. Plasmametadrenalines. Clin Sci (Colch). 1995;88:533-542.26. Eisenhofer G, Rundqvist B, Aneman A, et al. Re-gional release and removal of catecholaminesand ex-traneuronal metabolism to metanephrines.J Clin En-docrinol Metab. 1995;80:3009-3017.27. EisenhoferG, Keiser H, FribergP, et al.Plasmameta-nephrines are markers of pheochromocytoma pro-duced by catechol-O-methyltransferase within tu-mors. J Clin Endocrinol Metab. 1998;83:2175-2185.28. Eisenhofer G, Huynh T-T, Hiroi M, Pacak K. Un-derstanding catecholamine metabolism as a guide to

the biochemical diagnosisof pheochromocytoma. RevEndocrinol Metab Dis. 2001;2:297-311.29. Eisenhofer G, Aneman A, Hooper D, RundqvistB, FribergP. Mesentericorganproduction, hepaticme-tabolism,and renal eliminationof norepinephrine andits metabolites in humans. J Neurochem. 1996;66:1565-1573.30. Eisenhofer G. Free or total metanephrines for di-agnosis of pheochromocytoma: what is the differ-ence? Clin Chem. 2001;47:988-989.31. Walther MM,Reiter R, KeiserHR, et al.Clinicalandgenetic characterization of pheochromocytoma in vonHippel-Lindau families. J Urol. 1999;162:659-664.32. Pauker SG, Kopelman RI. Interpreting hoof-beats. N Engl J Med. 1992;327:1009-1013.33. Bravo EL, Tarazi RC, Fouad FM, Vidt DG, Gif-fordRW Jr. Clonidine-suppression test. N EnglJ Med.1981;305:623-626.34. GrossmanE, Goldstein DS, HoffmanA, Keiser HR.

Glucagon and clonidine testing in the diagnosis ofpheochromocytoma. Hypertension. 1991;17:733-741.35. PacakK, Linehan WM,EisenhoferG, Walther MM,Goldstein DS. Recent advances in genetics, diagno-sis, localization, and treatment of pheochromocy-toma. Ann Intern Med. 2001;134:315-329.36. Goldstein DS, Stull R, Markey SP, Marks ES, Kei-ser HR. Dihydrocaffeic acid: a common contaminantin the liquid chromatographic-electrochemical mea-surement of plasma catecholamines in man. J Chro-matogr. 1984;311:148-153.



Documents

Biochemical Diagnosis of Pheochromocytoma- Which Test is Best.full