Radioimmunoassay for plasma histamine: a study of false positive and false negative values

Transcription

1 British Journal of Anaesthesia 1995; 74: LABORATORY INVSTIATIONS Radioimmunoassay for plasma histamine: a study of false positive and false negative values D. LAROCH, F. DUBOIS, J.-L. RARD, C. LFRANCOIS, B. ANDR, M.-C. VRNAUD, L. DUBUS AND H. BRICARD Summary In order to ahieve a orret diagnosis of drug anaphylaxis using a radioimmunoassay devoid of interferenes, we have studied fators leading to false positive or false negative values of plasma histamine. Different steps in sample olletion were studied systematially in 30 normal volunteers. False positive values were found in haemolysed samples, with histamine onentrations being orrelated with haemoglobin onentrations, and where plasma was aspirated from the white-ell layer. There was no signifiant inrease when a tourniquet or vauum tubes were used, or when blood tubes were left at 4 C overnight. In 12 patients who experiened an anaphylati reation, histamine disappeared from blood 10 times more slowly than expeted. False negative values were found in two pregnant women and one heparinized patient. Histamine was remarkably stable in vitro in blood or plasma samples, whereas it disappeared rapidly when plasma from a pregnant woman or a heparinized patient was added to the sample. We onlude that false positive and false negative values are rare when using this radioimmunoassay. (Br. J. Anaesth. 1995; 74: ) Key words Histamine. Measurement tehniques, radioimmunoassay. There are many reasons for adverse reations to drugs. Anaphylati reations are of partiular interest as their severity inreases with repeated administration of the drug responsible. Biohemial diagnosis may be obtained by measuring plasma histamine onentrations soon after the reation [1,2]. However, onentrations were rarely measured in routine pratie, for two reasons: sampling reommendations [3] were diffiult to asertain in emergeny onditions and false positive results may be obtained when basophils are disrupted during or after sample withdrawal; and the half-life of histamine in plasma is assumed to be very short (minutes) [4] beause of metabolization by two pathways, methylation by methyltransferase and oxidation by diamine oxidase [5]. Thus falsenegative values may be obtained if blood is obtained late after the reation or if histamine is metabolized in the sample itself. The assays for histamine determination during the 1980s often laked sensitivity or speifiity [6], and ontroversial results were obtained in a quality ontrol study [7]. A radioimmunoassay, using a suinyl-glyinamide derivatization of histamine, has been available for a few years [8]. This assay is easy to use, sensitive and speifi [9], and appears to be reliable in linial onditions [10, 11]. The sampling reommendations, assessed for the lassi fluorimetri assay, may not apply when using the radioimmunoassay. Furthermore, in vivo metabolism of histamine under linial onditions, that is unexpeted anaphylati reations, has only been evaluated in a few patients [12,13] and the 1-min half-life found in volunteers perfused with histamine [4] may not be valid under suh onditions. The aim of this study was to determine whih onditions for sample olletion or handling ould lead to false positive or false negative values for plasma histamine onentrations when using the radioimmunoassay. Materials and methods CONTROL SUBJCTS AND SAMPLIN PROCDUR The study was approved by the loal Committee for the Protetion of Volunteers for Biomedial Researh and informed onsent was obtained from all subjets. Blood was obtained onseutively in seven tubes, within 1 or 2 h after lunh, from eah of 30 healthy volunteers (12 male) aged 27.3 (range 18-45) yr. On the left forearm, a tight tourniquet was plaed 5 min before introduing a 22-gauge Vautainer needle into an anteubital vein, and onneting it to 2-ml glass vauum tubes (Vautainer, Beton Dikinson). These tubes ontained, in order: tube 1, ethylenediaminetetraaeti aid (DTA); tube 2, DTA; tube 3, heparin; last tube, alled No. 7, DTA. At the same time a fluoroethylene-propylene atheter (Abboath T 20) was plaed, without a tourniquet, DOMINIQU LAROCH*, MD, PHD (Servie of Nulear Mediine); FRDRIC DUBOIS, MD, JAN-LOUIS RARD, MD, CLAUD LFRANCOIS, MD, BRNARD ANDR, MD, LAURNT DUBUS, MD, HNRI BRICARD, MD (Department of Anaesthesiology); MARI- CLAUD VRNAUD, MD (Department of Pneumonology); Centre Hospitalier Regional Universitaire, Caen, Frane. Aepted for publiation: November 5, *Address for orrespondene: Servie de Medeine Nuleaire, CHRU de Caen, F Caen Cedex, Frane.

2 Histamine: false positive and false negative values 431 into a vein of the right forearm and blood was obtained by gravity through the atheter into three DTA-ontaining open tubes, numbered onseutively 4 to 6 (inlusive). Tube Nos 4 and 6 were onstruted from polypropylene and tube 5 from sterile glass (Vautainer, Beton Dikinson). The whole sampling proedure took less than 5 min for eah subjet. Tube Nos 1-6 were plaed immediately in ie-old water and entrifuged at 4 C and 800 g (1500 rpm) for 10 min. Thereafter, plasma was aspirated gently without approahing the white ell layer and frozen at 20 C until assayed. Subsequently, from tube No. 2, another plasma sample was aspirated near or inside the white ell layer (tube No. 8) and frozen at 20 C. During the ourse of the study, spontaneous haemolysis ourred in three samples from two subjets and atheterization failed in two others. Tube Nos 1-6, but not tube No. 7 obtained from these four subjets were rejeted. Whole blood in tube No. 7 was left either at room temperature (20 C) for 1 (n = 8) or 18 h («= 10), or at 4 C for 18 («= 9) or 40 h (n = 7) before entrifugation. Furthermore, blood samples obtained from seven other volunteers were divided into three aliquots whih were entrifuged at either 4 C or 15 C or without refrigerating the entrifuge. SAMPLS WITH INCRASD HISTAMIN CONCNTRATIONS Two types of samples with inreased onentrations of histamine were used: plasma samples obtained in DTA-ontaining tubes from patients who had anaphylatoid reations, whih were stored frozen at 20 C, and blood samples from normal volunteers hallenged in vitro with anti-ig immunoglobulin (approximately 8 ml of blood was obtained in heparinized tubes (Vautainer, Beton Dikinson) and inubated at 37 C. Subsequently, 50 nl of monolonal anti-human-ig immunoglobulin (Immunoteh) was added and inubation was prolonged for 30 min). Histamine onentrations in the plasma samples were nmol litre" 1. Stability of histamine in whole blood After in vitro degranulation, blood samples were separated into two aliquots. One was entrifuged immediately and the other was left for 2 h at room temperature («= 9) or overnight (18 h) at 4 C (n = 7). In another experiment, five blood samples were divided into three aliquots whih were entrifuged at 4 C, 15 C or without refrigeration. Stability of histamine in plasma Three aliquots of plasma, obtained after in vitro degranulation of blood from five volunteers, were plaed in polypropylene, polyethylene or sterile glass tubes. The tubes were left for 30 min at room temperature, then frozen at 20 C until assayed. Plasma from three patients who experiened anaphylatoid reations and plasma obtained after in vitro degranulation were left in polypropylene tubes at room temperature for up to 48 h or at 4 C for up to 72 h, before measurement of histamine. Four plasma samples obtained during anaphylatoid reations were subjeted to five freeze-thaw yles, with measurement of histamine after eah yle. Seven samples obtained during anaphylatoid reations were diluted 1/10 or 1/50 in kit buffer before assay. The remaining volume of diluted plasma was frozen at 20 C and re-assayed 1 week later. IN VIVO HISTAMIN MTABOLISM Patients with inreased histamine onentrations Two onseutive blood samples were obtained in DTA-ontaining tubes from 12 patients (six male), aged yr, during anaphylatoid reations for measurement of plasma histamine and tryptase. All had immediately reated after i.v. administration of a drug (neuromusular bloker, n 8; beta-latam, n = 11 gelatine, n = 1; protamine, n = 1; aprotinin, n = 1). Ten patients reovered well but two who had experiened severe and prolonged anoxia died some weeks later. IN VITRO HISTAMIN MTABOLISM In order to evaluate the rate of histamine metabolism in the sample tube during speifi linial onditions, a plasma sample ontaining histamine 300 nmol litre" 1, obtained through in vitro degranulation, was mixed (vol/vol) with plasma from a pregnant woman (34 gestational weeks), from a heparinized patient (after 40 min of extraorporeal irulation) or from a normal ontrol. The mix was inubated at 37 C for 80 min or at 4 C for 22 h. ASSAYS Histamine was measured after alkylation by radioimmunoassay [8] (Immunoteh, Luminy, Frane). In this study, the lower limit of detetion of the assay was 0.5 nmol litre" 1 (1 nmol litre" 1 = 0.11 ng ml" 1 ). The usual limit for pathologial values is 9 nmol litre" 1 [3]. The within-assay oeffiient of variation is 10% for a plasma sample ontaining histamine 10 nmol litre" 1. For eah ontrol subjet, the eight tubes were proessed within the same assay series. In haemolysed samples, haemoglobin was measured by differential spetrosopy (normal range < 5 umol litre" 1 ) [14]. Tryptase was measured by an immunoradiometri assay [15] and urinary methylhistamine by a ompetitive radioimmunoassay [16] (Kabi Pharmaia Diagnostis, Uppsala, Sweden). The normal range is < 2 ng litre" 1 for tryptase and < 380 (imol per mole of reatinine for urinary methylhistamine [17]. All measurements were made in dupliate. When neessary, samples were diluted in the buffer provided with the kit. STATISTICAL ANALYSIS Data are expressed as mean (SD). Means were ompared by analysis of variane. Values were ompared by Wiloxon or Kruskal-Wallis tests when

3 432 British Journal of Anaesthesia the number of ompared samples was small. Linear orrelation was performed using the method of least squares. P < 0.05 was onsidered statistially signifiant. 10 Results SARCHIN FOR FALS POSITIV RSULTS IN NORMAL CONTROLS Blood sampling Sampling onditions are summarized in table 1. Individual histamine onentrations measured in the 30 volunteers are shown in figure 1, aording to the different sampling onditions. Mean histamine onentrations in tubes entrifuged immediately were, respetively (n = 30): tube 1, 2.04(1.02) (range ) nmol litre" 1 ; tube 2, 1.80 (0.80) ( ) nmol litre" 1 ; tube 3, 2.04 (0.72) ( ) nmol litre" 1 ; tube 4, 2.06(1.11) ( ) Table 1 Summary of the different withdrawal onditions used for obtaining blood samples in 30 volunteers for measurement of plasma onentrations of histamine. = ondition fulfilled, = glass, P = polypropylene, H = heparin; = DTA. *Tube left at room temperature or 4 C for various times Sampling ondition First ml disarded Tourniquet Catheter Vauum Tube material Antioagulant 4 C Immediate entrifugation Tube number H 4 P 5 6 p 7 * C 15 C WR Figure 2 Histamine onentrations measured after entrifugation at 4 C, 15 C or without refrigeration (WR), of blood aliquots from seven different subjets. nmol litre" 1 ; tube 5, 1.90 (0.94) (0.4-5) nmol litre" 1 ; and tube 6, 2.19(0.91) ( ) nmol litre" 1. The overall mean was 2.01 (0.92) nmol litre" 1. Analysis of variane showed no signifiant differene between means. One of the patients (represented by filled triangles in fig. 1) had inreased values in tubes 4 and 5 (6.2 and 5 nmol litre" 1, respetively). The values in the other tubes were nmol litre" 1. Stability of histamine in whole blood Blood tube No. 7 was left at room temperature for 1 or 18 h, or at 4 C for 18 or 40 h. Histamine onentrations were ompared with onentrations in tube No. 2 entrifuged immediately from the same patient. There was no signifiant differene between histamine onentrations in tube Nos 7 and 2 exept for a slight inrease after 18 h at room temperature (2.67(1.07) vs 1.81 (1.02) nmol litre" 1 ; Wiloxon test, P < 0.01) = 6 - o 4-2- I I! A Tube number Figure 1 Individual histamine onentrations in 30 volunteers, aording to the sampling onditions 1-6 (see table 1). = pathologial range (9 nmol litre" 1 = 1 ng ml" 1 ). Centrifuge temperature Blood aliquots from seven subjets were entrifuged at different temperatures: mean histamine onentrations were 3.66 (1.64) nmol litre" 1 (range nmol litre" 1 ) at 4 C; 1.81(0.48) ( ) nmol litre" 1 at 15 C; and 2.41 (0.57) ( ) nmol litre" 1 without refrigeration (fig. 2). The differene was signifiant (P < 0.01, Kruskal-Wallis test). In one ase, a large inrease in plasma histamine onentration was observed at 4 C. When this patient was exluded, the differene was still signifiant (P<0.02); values found after entrifugation at 15 C were onsistently smaller than in the other onditions. Plasma aspiration Histamine onentrations exeeded 9 nmol litre" 1 in six of 30 samples obtained near the buffy oat (tube No. 8). The onentration range was 0.6- > 150 nmol litre" 1. The mean onentration was 12.3 (28.3) nmol litre" 1, whereas in plasma aspirated from the same tube far from the buffy oat (tube No.

4 Histamine: false positive and false negative values , 30H "o 20-0) <S 10- _ O _ <o Haemoglobin (nmol litre" 1 ) Figure 3 Linear orrelation between histamine and haemoglobin onentrations in eight spontaneously haemolysed plasma samples (r 2 = 0.992). 2), it was 1.8 (0.8) nmol litre" 1 (Wiloxon test, P < 0.05). The measured value was related to the level of plasma reahed by the pipette. The value remained unhanged when approahing the ell layer until the tip of the pipette entered the white ell layer, then the value inreased dramatially. Spontaneous haemolysis Two ontrol subjets were rejeted from the study beause haemolysis ourred, respetively, in 1 or 2 samples: histamine onentrations were 5.5, 12.4 and 11.4 nmol litre" 1, whereas the mean values in the other tubes from the same subjets were, respetively, 1.74(0.48) and 3.18 (0.70) nmol litre" 1. In order to evaluate the effet of haemolysis, we measured histamine and haemoglobin onentrations in 10 spontaneously haemolysed plasma samples obtained for other purposes. Histamine onentrations ranged from 2.1 to 32 nmol litre" 1 and haemoglobin from 3.2 to 49 umol litre" 1. Two samples had inreased haemoglobin (13.5 and 33.6 umol litre" 1, respetively) but normal histamine onentrations (2.8 and 5.5 nmol litre" 1 ). A linear orrelation (r 2 = 0.992) was found for the eight other pairs of values (fig. 3). SARCHIN FOR FALS NATIV RSULTS In vitro stability studies Stability of histamine in whole blood. We ompared histamine onentrations in blood aliquots entrifuged immediately after degranulation with onentrations in aliquots left for 2 h at room temperature or 18 h at 4 C before entrifugation. No signifiant differene was found between the values in tubes entrifuged immediately and in the other tubes (Wiloxon test). Similarly, the effet of temperature during entrifugation was studied at 4 C, 15 C or without refrigeration. The KruskalWallis test showed no signifiant differene between histamine onentrations in the three aliquots. Stability of histamine in plasma. The effet of storage of plasma in polypropylene, polyethylene or sterile 10 J Figure 4 ffet of freezing diluted plasma samples at 20 C on histamine onentrations (values orreted for dilution). 1 = Immediate assay, 2 = assay after freezing. glass tubes was studied. No signifiant differene was found for histamine onentrations in the different tubes (Kruskal-Wallis test). No signifiant differene was found between histamine onentrations measured immediately after entrifugation or after inubation of plasma at room temperature for up to 48 h, or at 4 C for up to 72 h (Kruskal-Wallis test). Five suessive freezethaw yles were performed for four plasma samples. There was no signifiant differene between the values after 1, 2, 3, 4 or 5 yles (Kruskal-Wallis test). Moreover, a plasma sample has been stored frozen at 20 C for 3 years, with histamine onentrations of 1116 nmol litre" 1 after the reation, 1060 nmol litre" 1 4 months later and 1047 nmol litre" 1 3 years later. Diluted plasma samples were assayed immediately after the dilution proess or frozen at 20 C before assay. The measured histamine onentrations were smaller in frozen diluted samples (Wiloxon test, P < 0.05); the differene was greater when the initial onentration was smaller (fig. 4). In vivo histamine metabolism Blood was obtained twie from 12 patients who experiened anaphylati reations. The seletion riteria were an inrease in histamine and tryptase onentrations during the reation. All had inreased plasma histamine onentrations in the first sample (853(1343) ( ) nmol litre" 1 ), obtained 5-60 min (30 (21) min) after the onset of the reation, and plasma tryptase onentrations were ug litre" 1. The mean histamine onentration in the seond sample was 65 (95) (5-300) nmol litre" 1, min (72 (43) min) after the first. As shown in figure 5, all but one patient had persistently inreased histamine onentrations 60 min after the onset of the reation. Clinial ases with false negative results Case 1. A pregnant woman, aged 35 yr, was administered suxamethonium and thiopentone for Caesarean setion. Immediately, erythema, hypotension and mild bronhospasm ourred. Blood was obtained 30 min after the onset of the reation. Plasma histamine onentration was < 0.5 nmol litre" 1, whereas plasma tryptase was 35 ug litre" 1 (normal

5 434 British Journal of Anaesthesia value < 2 ug litre" 1 )- A seond blood sample, obtained 1 h later, ontained histamine < 0.5 nmol litre" 1 and tryptase 47 ug litre" 1. Urine samples ontained methylhistamine 1180 nmol per mole of reatinine (normal range < 380 nmol mol" 1 ). ight weeks later, skin tests and speifi Ig measurements showed evidene of anaphylaxis to suxamethonium ~ 100- ~ 10- Case 2. A pregnant woman, aged 32 yr, had a Caesarean setion for a twin pregnany. She reeived suxamethonium and thiopentone and developed aute bronhospasm, followed by diffuse urtiaria and severe hypotension 10 min later. A blood sample was withdrawn immediately, but left at 4 C overnight. Plasma histamine onentration was 0.6 nmol litre" 1 and tryptase 63 ug litre" 1. In a seond sample, 1 h later, histamine onentration was 0.6 nmol litre" 1 and tryptase 86 ug litre" 1. Urinary methylhistamine onentration was 2160 umol per mole of reatinine. Neuromusular bloker-speifi Ig was present in high onentrations and skin tests were positive for suxamethonium. Case 3. A 56-yr-old man was administered fentanyl, fiunitrazepam, thiopentone, panuronium, Haemael and heparin before extraorporeal irulation (CC) for oronary bypass operation. When CC was started, aute hypotension was observed, whih was treated by aeleration of CC, ephedrine and fluid perfusion. A blood sample was obtained within 10 min of the reation. Histamine onentration was 1.5 nmol litre" 1 and tryptase 4.3 ug litre" 1. One hour later, histamine onentration was < 0.5 nmol litre" 1 and tryptase 9.4 ug litre" 1. Urinary methylhistamine onentration was 1665 umol per mole of reatinine. The moleule responsible was not identified. In vitro histamine metabolism We studied the rate of disappearane of histamine from a plasma sample ontaining histamine, 300 nmol litre" 1, mixed with plasma from a pregnant woman (mix No. 1), from a heparinized patient (mix No. 2) or from a normal subjet (mix No. 3). When 10000i o S Time (min) Figure 5 Histamine onentrations (log) in two suessive samples from 12 patients who experiened anaphylatoid reations Time (min) Figure 6 In vitro disappearane of histamine at 37 C from plasma ontaining histamine 300 nmol litre" 1, mixed (vol/vol) with normal plasma ( ), plasma from a pregnant woman (D) or plasma obtained after 40 min of extraorporeal irulation (O). inubation was performed at 37 C, histamine disappeared from mix Nos 1 and 2 with an apparent half-life of 20 min (fig. 6). At 4 C, the half-life was 130 min for mix No. 1 and 110 min for mix No. 2. After 22 h at 4 C, histamine onentrations were 4.2 nmol litre" 1 in mix No. 1 and 1.8 nmol litre" 1 in mix No. 2. Histamine onentrations did not hange in mix No. 3 during the whole experiment. Disussion The aim of this study was to determine whih onditions ould lead to false positive or false negative values for plasma onentrations of histamine when using the radioimmunoassay, in order to ahieve an aurate diagnosis of anaphylaxis. The monolonal antibody used here is highly speifi for histamine; the ross-reativity ratio for N-methylhistamine is 1/14500 and for histidine 1/ [8]. Numerous ompounds known to be potentially apable of interfering with measurements of histamine failed to interfere in this assay [9]. This assay ompares favourably with other tehniques, aording to a quality ontrol study [18]. It was shown to have equal sensitivity and auray as the fluorimetri assay, and to be onvenient to use for plasma samples at the Munih Consensus Development Conferene on Histamine Determination [19]. Aording to the manufaturer, the sensitivity is 0.2 nmol litre" 1. In this study it was 0.5 nmol litre" 1 (78 determinations over a 2-yr period). The variation oeffiient was 10% for a onentration of 10 nmol litre" 1 (21 determinations) and in the normal onentration range [9]. False positive results were investigated in normal volunteers after lunh, as food may inrease plasma histamine onentrations. The mean histamine onentration was 2.01 (0.92) nmol litre" 1, whereas it was 0.8 (0.4) nmol litre" 1 in 13 fasting ontrols [10] (P < 0.01). Other laboratories reported values of 3.33 (1.62) nmol litre" 1 [8] in 14 non-allergi subjets or 1.74 (0.72) nmol litre" 1 [9] in 40 normals using the same reagents. From our results we suggest that a simplified sampling proedure may be used omprising diret venous punture and sampling into glass vauum tubes ontaining either DTA or heparin. The

6 Histamine: false positive and false negative values 435 threshold usually reported for pathologial levels, obtained with the fluorimetri method, is 1 ng ml" 1, that is 9 nmol litre" 1 [3]. This value appears to be appropriate for the radioimmunoassay also. The stability of basophils in blood has been studied at room temperature and at 4 C, and during entrifugation at various temperatures. Although there was a tendeny for histamine to inrease when blood was left overnight at room temperature or for 2 days at 4 C, the samples ould be left for 1 h at room temperature or for 1 night at 4 C without signifiant hanges in histamine onentrations, indiating that histamine does not leak from basophils readily enough to give false positive values. The values obtained after entrifugation at different temperatures were very lose, exept when the entrifugation temperature was 4 C. This ould be explained by poor temperature regulation, with variations of + 2 C or 3 C, and potential freezing leading to ell lysis. Thus we suggest die use of an intermediate temperature of 15 C. Although use of a nonrefrigerated entrifuge is possible, it indues a slight inrease in histamine onentrations. As entrifugation was performed at 4 C for tubes 1-6, it is possible that the higher values obtained apparently at random in some samples ould result from an exessively low temperature. During this study, false positive values were observed in two different situations: (i) when spontaneous haemolysis ourred (three of 224 samples, i.e. 1.2 %), histamine onentrations ranged from normal to moderately inreased. Histamine inreased onomitantly with an inrease in haemoglobin in most ases. Spontaneous haemolysis during sampling affets a few ells: haemoglobin onentrations are in the miromolar range in haemolysed plasma and millimolar in whole blood, and histamine onentrations in haemolysed plasma are times smaller than in whole blood. However, the red olour is readily notied. Measurements of haemoglobin may be useful in interpreting an inreased histamine onentration in a reddish sample obtained after an anaphylatoid event; (ii) when plasma was aspirated from the upper ell layer, a dramati inrease in histamine onentration ourred. In suh ases, no apparent hange was notied in the tested plasma, ontrary to the haemolysed samples. We were also interested in deteting possible false negative results aused by the disappearane of histamine either in vivo or in vitro. The in vivo halflife of plasma histamine has been alulated as 1-2 min in volunteers reeiving histamine infusions [4] and was shown to be in the minute range during hallenges [20] or protools [21]. All of these values were obtained in subjets with moderately inreased peak histamine onentrations. We seleted for this study 12 patients with inreased onentrations of plasma histamine during anaphylati reations to i.v. administered drugs. All had inreased onentrations of tryptase, whih is a speifi marker of mast ell ativation virtually absent from normal serum [22]. The initial histamine onentrations were inreased markedly in the majority of patients ( nmol litre" 1 ). As no intermediate values were obtained, the exat half-life ould not be alulated; however its value was probably loser to 20 than to 2 min. The longer half-life ould be explained by saturation of enzymati metabolism, by ontinued release or by redued learane beause of the disease proess. The ombination of a high initial histamine onentration and a longer half-life than expeted allowed histamine to remain at pathologial levels 1 h after the onset of reation in 11 of 12 patients. The stability of histamine was studied further in vitro. As large quantities of different plasma obtained during anaphylatoid reations were not easily available, in some experiments we used normal blood hallenged in vitro with anti-ig. On suh oasions the antioagulant was heparin instead of DTA. However, heparin has no effet in vitro on histamine metabolism and we have verified in one patient that histamine onentrations were idential in DTA and heparin-ontaining tubes obtained at the same time after an anaphylatoid event. Whole blood ould be left for 2 h at room temperature or overnight at 4 C without signifiant hange in histamine onentration. Blood ould be entrifuged at 4 C, 15 C or without refrigeration, without hanging histamine values. It had been suggested that histamine ould be adsorbed on tube walls, espeially where sterile glass is used, and that some ontaminants of the tube material ould interfere with the assay [23], but we ould not onfirm this. It is likely that ontaminants interfered with light emission during the fluorimetri assays, whereas radioimmunoassays are not affeted by suh onditions, as they are highly speifi for the measured moleule. Dyer and olleagues reported that histamine added to normal plasma was stable for 30 min at room temperature or at 4 C, but did not experiment further [24]. In this study, endogenous histamine was stable in separated plasma for 48 h at room temperature or 72 h at 4 C. It is usual to suggest that one should not re-freeze plasma samples that have previously been frozen. However, five freeze-thaw yles did not hange histamine onentrations in this study. When diluted plasma samples were frozen, a dramati derease in histamine onentrations was observed. The radioimmunoassay used in this report proeeds through an alkylation step leading to stable alkylated histamine [8]. When diluted histamine samples were alkylated before freezing, there was no derease in histamine onentration (results not shown). Thus, when in vitro histamine-release tests are performed from diluted blood samples, supernatants should be either assayed immediately or alkylated before freezing if the assay annot be performed immediately. False negative values have been identified in two different linial onditions: pregnany and extraorporeal irulation. Tryptase onentrations are inreased in patients undergoing anaphylatoid reations [10,25]. Thus the finding of inreased onentrations of tryptase in eah of our three patients indiated mast ell degranulation, and obviously in vivo histamine release, whih was onfirmed by inreased methylhistamine onen-

7 436 British Journal of Anaesthesia trations in urine. There are two reasons for the low value of histamine: a shorter half-life in vivo and in vitro degradation. Several reports have shown that diamine oxidase ativity inreases during pregnany, with maximum ativation during the third gestational trimester [26, 27]. It inreases to similar levels 30 min after injetion of large doses of heparin [28]. In patients administered low doses of heparin, the effet appears to be dose-dependent [29]. We tried to appreiate the in vivo atabolism of histamine under similar linial onditions: mixing plasma ontaining large quantities of endogenous histamine with plasma from a pregnant woman or a heparinized patient led to the disappearane of histamine from the sample within 80 min at 37 C, or 1 night at 4 C, whereas histamine was remarkably stable in plasma from subjets with normal metabolism. Our results onfirm those of Morel and Delaage who showed that exogenous histamine added to plasma from a pregnant woman disappeared from the sample within 10 min whereas alkylated histamine did not [8]. These findings indiate that preautions should be taken in order to avoid false negative values for histamine in pregnant women and heparinized patients: blood should be refrigerated and plasma frozen as soon as possible, preferably after alkylation of histamine. In ontrast, patients with normal metabolism of histamine do not require suh preautions, as histamine is remarkably stable in their plasma samples. This study has shown that a great number of preautions used with the fluorimetri assay are not neessary with the radioimmunoassay, and that false positive results are rare. Moderately inreased onentrations (< 50 nmol litre" 1 ) may be found in haemolysed samples. High values are found where plasma is aspirated from the white ell layer. Thus laboratory staff must be well informed of the dramati onsequenes in subsequent diagnosis. During anaphylati events, histamine disappeared from plasma at a slower rate than expeted, and high values were found 1 h, and in some patients 2 h, after the onset of the reation. This allows emergeny treatment to be administered before sampling is performed. Clinial onditions where histamine metabolism is ativated are rare. They inlude pregnany and high-dose heparinization. In suh ases false negative values are likely to our. Plasma tryptase should be measured in order to ensure an aurate diagnosis, although the linial sensitivity of this marker is lower than that of histamine [17]. Histamine measurements should be undertaken every time a linial reation suggesting anaphylaxis ours after drug administration. When resusitation is unsuessful or after-effets are antiipated, suh measurements may give forensi evidene [11]. Where the reation mehanism is unlear, they may be helpful. Furthermore, during drug or food hallenges they an be of great interest for linking together mehanisms and symptoms. As routine, reliable tehniques are available, and as false negative, and also false positive, results are rare and well doumented, there is no justifiation, apart from ost, for not performing histamine measurements every time drug anaphylaxis is suspeted. Aknowledgements This work was supported by a grant from the National Health Ministry. The anaesthetists who partiipated in the study are kindly aknowledged. We thank the Laboratory of Nulear Mediine for exellent tehnial assistane, the Laboratory of Haemostasis for Cardiovasular Diseases (M.-C. Khayat, MD) for performing the haemoglobin measurements and Immunoteh International for providing the histamine kits. Referenes 1. Bohner BS, Lihtenstein LM. Anaphylaxis. New ngland Journal of Mediine 1991; 324: Wasserman SI. Mediators of immediate hypersensitivity. Journal of Allergy and Clinial Immunology 1983;72: Lorenz W, Neugebaueur, Shmal A. Plasma histamine assay for anaphylatoid reations in the anaesthetized subjet. ffets of blood olletion and plasma preparation on the measured histamine. Annales Franoises a"anesthesie et de Reanimation 1982; 1: Ind PW, Brown MJ, Lhoste FJM, Maquin I, Dollery CT. Determination of histamine and its metabolites. Conentration effet relationships of infused histamine in normal volunteers. Agents and Ations 1982; 12: Shayer RW. Catabolism of histamine in vivo. Handbook of xperimental Pharmaology 1966; 18: Ind PW, Barnes PJ, Brown MJ, Causon R, Dollery CT. Measurement of plasma histamine in asthma. Clinial Allergy 1983; 13: leih J, Hull WM. Measurement of histamine: a quality ontrol study. Journal of Allergy and Clinial Immunology 1980; 66: Morel AM, Delaage MA. Immunoanalysis of histamine through a novel hemial derivatization. Journal of Allergy and Clinial Immunology 1988; 82: MBride P, Bradley D, Kaliner M. valuation of a radioimmunoassay for histamine measurement in biologi fluids. Journal of Allergy and Clinial Immunology 1988; 82: Larohe D, Vergnaud MC, Sillard B, Soufarapis H, Briard H. Biohemial markers of anaphylatoid reations to drugs. Comparison of plasma histamine and tryptase. Anesthesiology 1991; 75: Larohe D, Lefranois C, erard JL, Dubois F, Vergnaud MC, ueant JL, Briard H. arly diagnosis of anaphylati reations to neuromusular bloking drugs. British Journal of Anaesthesia 1992; 69: Moss J, Fahmy NR, Sunder N, Beaven MA. Hormonal and hemodynami profile of an anaphylati reation in man. Cirulation 1981; 63: Smith PL, Kagey-Sobotka A, Bleeker R, Traystman R, Kaplan AP, ralnik H, Valentine MD, Permutt S, Lihtenstein LM. Physiologi manifestations of human anaphylaxis. Journal of Clinial Investigation 1980; 66: Cripps CM. Rapid method for the estimation of plasma haemoglobin level. Journal of Clinial Pathology 1968; 21: nander I, Matsson P, Nystrand J, Andersson AS, klund, Bradfort TR, Shwartz LB. A new radioimmunoassay for human mast ell tryptase using monolonal antibodies. Journal of Immunologial Methods 1991; 138: Adriani, De Petrillo, Bonini S. A new radioimmunoassay for histamine determination. Allergy 1988; 43 (Suppl. 7): Larohe D, Dubois F, Lefranois C, Vergnaud MC, erard JL, Soufarapis H, Sillard B, Briard H. arly biologial markers of anaphylatoid reations ourring during anaesthesia. Annales Franoises <f Anesthesie et de Reanimation 1992; 11: Oosting, Neugebauer, Keyzer JJ, Lorenz W. Determination of histamine in human plasma: the uropean xternal Quality Control Study Clinial xperimental Allergy 1990; 20: Lorenz W, Neugebauer, Uvnas B, Beaven MA, nnis M, ranerus, reen JP, Keyzer JJ, MBride PT, Mannaioni PF, Peare FL, Watkins J. Munih onsensus development onferene on histamine determination. Handbook of xperimental Pharmaology 1991; 97:

8 Histamine: false positive and false negative values Kaplan AP, ray L, Shaff R, Horakova Z, Beaven MA. In vivo studies of mediator release in old urtiaria and holinergi urtiaria. Journal of Allergy and Clinial Immunology 1975; 55: 394-^ Moss J, Rosow C, Savarese JJ, Philbin DM, Kniffen KJ. Role of histamine in the hypotensive ation of d-tubourarine in humans. Anesthesiology 1981; 55: Castells M, Irani AA, Shwartz LB. valuation of human peripheral blood leukoytes for mast ell tryptase. Journal of Immunology 1987; 138: Lorenz W, Doenike A. Histamine release in linial onditions. Mount Sinai Journal of Mediine (NY) 1978; 45: Dyer J, Warren K, Merlin S, Metalfe DD, Kaliner M. Measurement of plasma histamine: desription of an improved method and normal values. Journal of Allergy and Clinial Immunology 1982; 70: Shwartz LB, Metalfe DD, Miller JS, arl H, Sullivan T. Tryptase levels as an indiator of mast ell ativation in systemi anaphylaxis and mastoytosis. New ngland Journal of Mediine 1987; 316: Ahlmark A. The histaminolyti power of plasma in man. Ada Physiologia Sandinavia 1944; 9 (Suppl. 28): Torok, Brewer JI, Dolkart R. Serum diamine oxidase in pregnany and in trophoblasti diseases. Journal of Clinial ndorinology and Metabolism 1970; 30: Hansson R, Holmberg C, Tibbling, Tryding N, Westling H, Wetterquist H. Heparin-indued diamine oxidase inrease in human blood plasma. Ada Media Sandinavia 1966; 180: Baylin SB, Beaven MA, Krauss RM, Keiser HR. Response of plasma histaminase ativity to small doses of heparin in normal subjets and patients with hyperlipoproteinemia. Journal of Clinial Investigation 1973; 52: