Gut 1996; 39: 13-17 Dereased plasma isoleuine onentrations after upper gastrointestinal haemorrhage in humans 13 Department of Surgery, Biomedial entre/aademi Hospital Maastriht, Maastriht, the Netherlands H Dejong W J H J Meijerink L H van Berlo N E P Deutz P B Soeters orrespondene to: Dr H Dejong, Department of Surgery, De Wever Hospital Heerlen, PO Box 4446, NL-641 X Heerlen, the Netherlands. Aepted for publiation 31 January 1996 H Dejong, W J H J Meijerink, L H van Berlo, N E P Deutz, P B Soeters Abstrat Bakground-A derease in arterial isoleuine values after intragastri blood administration in pigs has been observed. This ontrasted with inreased values of most other amino aids, ammonia, and urea. After an isonitrogenous ontrol meal in these pigs all amino aids inluding isoleuine inreased, and urea inreased to a lesser extent, suggesting a relation between the arterial isoleuine derease and uraemia after gastrointestinal haemorrhage. Methods-To extend these findings to humans, plasma amino aids were determined after gastrointestinal haemorrhage in patients with pepti ulers (n=9) or oesophageal varies indued by liver irrhosis (n=4) and ompared with preoperative patients (n= 16). Results-After gastrointestinal haemorrhage, isoleuine dereased in all patients by more than 6% and normalised within 48 hours. Most other amino aids inreased and also normalised within 48 hours. Uraemia ourred in both groups, hyperammonaemia was seen in patients with liver irrhosis. onlusions-these results onfirm previous findings in animals and healthy volunteers that plasma isoleuine dereases after simulated upper gastrointestinal haemorrhage. This supports the hypothesis that the absene of isoleuine in blood protein auses dereased plasma isoleuine values after gastrointestinal haemorrhage, and may be a ontributory fator to uraemia and hyperammonaemia in patients with normal and impaired liver funtion, respetively. Intravenous isoleuine administration after gastrointestinal haemorrhage ould be benefiial and will be the subjet of further researh. (Gut 1996; 39: 13-17) Keywords: human, amino aids, gastrointestinal haemorrhage, irrhosis, pepti uler disease, ammonia, urea. Upper gastrointestinal haemorrhage indues profound malaise, espeially in patients with impaired liver funtion. 1 This annot be explained solely by the blood loss, as blood donors or even trauma patients, losing onsiderable amounts of blood, do not experiene the harateristi malaise ofpatients with upper gastrointestinal bleeding. Apparently, the fat that the blood is shed into the digestive trat plays an important part in the pathophysiology and linial piture of upper gastrointestinal haemorrhage in patients with and without liver disease. Patients with impaired liver funtion often develop hepati enephalopathy after gastrointestinal haemorrhage.' This has been related to rising ammonia onentrations,l probably aused by the inreased nitrogen load in the presene of portasystemi shunting and liver disease, with resulting inadequate ammonia removal through urea synthesis.2 After administration of equal amounts ofblood, patients with liver disease show a far greater rise in blood ammonia onentrations than patients with normal liver funtion.1 In this ontext, it has been suggested that blood is more ammoniageni than other nitrogen soures in animals and patients with or without liver irrhosis. 1 3 Several explanations have been put forward for these findings, but none of them fully aounts for the 'toxi' effets of blood in the digestive trat. In reent years, we have attempted to eluidate the ammoniageni potential of blood in the digestive trat, by studying the effets of blood ingestion on amino aid and ammonia metabolism in pigs and healthy volunteers.3-5 Within six hours after intragastri blood administration in healthy pigs as well as after a blood pudding meal in healthy volunteers, a more than 5%/ derease in the plasma values of the essential branhed hain amino aid isoleuine (ILE) was observed, while the onentration of the other two branhed hain amino aids leuine (LEU) and valine (VAL) and most other amino aids reahed extremely high onentrations.3-5 Simultaneously, arterial urea and ammonia values inreased.3 Isonitrogenous ontrol meals led to an inrease of most amino aids, inluding ILE, and a less pronouned inrease in arterial ammonia and urea. Analysis of the amino aid omposition of blood protein (mainly haemoglobin) showed an almost omplete absene of the essential amino aid ILE.3 As a onsequene, blood protein has a low biologial value.6 Thus, we hypothesised that the near absene of ILE in blood protein was ausally related to the diminished plasma ILE onentrations as well as to the uraemia and hyperammonaemia after simulated gastrointestinal haemorrhage in experimental animals. As a logial next step, we verified this hypothesis by supplementing ILE intravenously simultaneously with intragastri blood administration in our pig model.4 The derease in arterial WE ould be prevented and the rise in arterial urea and ammonia values ould be blunted by simultaneous intravenous ILE supplementation in these pigs, suggesting improved protein and amino aid utilisation.4 Gut: first published as 1.1136/gut.39.1.13 on 1 July 1996. Downloaded from http://gut.bmj.om/ on 15 January 219 by guest. Proteted by opyright.
14 TABLE I Patient data These findings in experiments performed in otherwise healthy animals and human volunteers35 ould be of potential linial relevane. If similar hanges were seen after upper gastrointestinal haemorrhage in humans, this would lend support for the supplementation of ILE during upper gastrointestinal bleeding in linial pratie. This study was undertaken to further explore the appliability of this hypothesis. As a first step, we investigated whether plasma ILE also dereases after endosopially verified upper gastrointestinal bleeding in otherwise healthy humans. In addition, four patients with liver irrhosis were studied. Methods Patients Thirteen patients with major, endosopially verified upper gastrointestinal bleeding from pepti duodenal or gastri ulers (n=9; mean (SD) age 65 (7) years (range 23-84)) or from oesophageal varies resulting from liver irrhosis (n=4; mean age 43 (4) years (range 33-5)) were studied. The most important exlusion riteria were malignanies (for metaboli reasons) and, for ethial reasons, pre-existent ardiovasular disease, irulatory instability, and primary oagulation disorders. Table I lists the patient harateristis. Written informed onsent was obtained from the patient or the patient's next of kin. The projet was approved by the loal institutional medial ethial ommittee. Blood sampling and determinations Blood for ammonia, urea, and amino aid analysis was sampled in the fasted state at the time points indiated in the Figures. The shedule was designed to sample hourly during the first six hours (starting immediately after admittane to the hospital), two hourly during the next six hours, four hourly until 24 hours, and then every 12 hours until 12 hours. If possible, a final sample was obtained 24 hours after admission. The presumed start of the bleeding, assessed by areful history taking, was taken as t= hours in all ases. Anteubital venous blood was sampled from the arm opposite to the infusion site with a vauum system (Venajet System) in 5 ml sodiumheparin tubes (ammonia blank less than 2,uM). Samples were kept on ie during transport to the laboratory and were treated, Patient Disease hild's lass Sex Age A Duodenal uler male 84 B Gastri uler male 84 Gastri uler female 51 D Duodenal and gastri uler female 79 E Gastri uler male 73 F Gastri uler male 23 G Duodenal uler male 76 H Gastri uler male 54 I Gastri uler male 64 J Aloholi irrhosis B/ male 5 K Aloholi irrhosis male 45 L Aloholi irrhosis male 33 Intravenous drug addition M ryptogeni irrhosis A male 44 Diabetes mellitus. Dejong, Meijerink, van Berlo, Deutz, Soeters stored at - 7, and determined as detailed reently.47 8 Ammonia and urea were determined enzymatially and amino aids by high performane liquid hromatography tehnique in plasma.47 8 The oeffiient of variane for all determinations was less than 4%. Statistis For statistial analysis, the total sample period was divided in six intervals (-12 h, 12-24 h, 24-48 h, 48-72 h, 72-96 h, 96-24 h). For eah time period, the mean of all samples per patient was alulated and this mean was used for analysis. Thus, eah patient generated a maximum of six data. One way analysis of variane was used to test for main effets of time within eah group. Mann-Whitney U nonparametri testing was used for omparison of eah time period with a ontrol group of preoperative eletive surgial patients (ontrols; n= 16), studied simultaneously also for other purposes detailed elsewhere.8 A p< 5 level was onsidered signifiant. Results General In most patients, the delay between the start of the bleeding and the first sample (estimated bleeding time) was approximately six hours. Pepti uler patients (Figs 1 and 2, Table II) At six hours after upper gastrointestinal bleeding, ILE was only approximately 3-4% of ontrol values in all uler patients (Fig 1; _I-----m ) Q -) 2 25 2 F VAL + LE:U =. 15 '-r X 1 ) r 5 U. 5 / 5 2 25 Time (hi) Figure 1: Plasma ILE (upper panel) and the sum of VAL and LEU (lower panel) onentrations after major, endosopially verified, upper gastrointestinal haemorrhage in pepti uler patients (n =9; mean age 65years (range 23-84)). Data are means (SEM) (MM). The start of bleeding, assessed by areful history taking, is taken as t= hours. The horizontal dashed line onnetingfilled retangles represents normal ontrol values ofplasma ILE obtained in a group of 16 eletive preoperative surgial patients. Gut: first published as 1.1136/gut.39.1.13 on 1 July 1996. Downloaded from http://gut.bmj.om/ on 15 January 219 by guest. Proteted by opyright.
Plasma isoleuine after gastrointestinal bleeding 15 i -.h 4).) 2 25 Figure 2: The aromati amino aids (AAA (tyrosine +phenylalanine); upper panel) and the sum of the individual amino aids measured (SUMAA, lower panel) onentrations in plasma after major, endosopially verified, upper gastrointestinal haemorrhage in pepti uker patients (n=9; mean age 65 years (range 23-84)). Data are means (SEM) (MM). The start of bleeding, assessed by areful history taking, is taken as t= hours. The horizontal dashed line onnetingfilled retangles represents normal ontrol values ofplasma ILE obtained in a group of 16 eletive preoperative surgial patients. TABLE II AAA 5 / p<q1). ILE inreased with time (p< 5) and returned within 48 hours to values not signifiantly different from ontrols (p> 5). The other two branhed hain amino aids, VAL and LEU, showed the opposite pattern: they were inreased after upper gastrointestinal bleeding and normalised within 24-48 hours (Fig 1; p< 1). The aromati amino aids (Fig 2) behaved idential to VAL and LEU; they were inreased after upper gastrointestinal haemorrhage ompared with ontrols and returned to onstant values within 24-48 hours (p<1). A similar pattern was seen for the sum of individual amino aids measured: there was a signifiant effet of time (p< 5). For the remaining amino aids, ammonia, and urea (Table II) the following patterns were seen: MET, ARG, and GLY showed a transient derease of maximally 25% (ILE pattern). TRP values did not hange in time 2 25 but were dereased ompared with ontrols. Some amino aids remained essentially unhanged (IT, GLU, ASN, SER, THR, LYS). Other amino aids (GLN, HIS, ALA) were initially similar to ontrol values, but dereased to stable values signifiantly lower than ontrols. The remainder (alaba, TAU) followed a pattern similar to the aromati amino aids. Ammonia remained unhanged and urea values were strongly inreased and only slowly returned to normal. irrhoti patients (Figs 3 and 4) ILE was dereased after gastrointestinal haemorrhage in irrhoti patients and normalised within 48-72 hours. Despite the small number of patients, this derease in ILE values was signifiant in the first 12 hours (p<5, nonparametri). In subjet K massive haemorrhage ourred in the hospital during injetion of slerosing agents for bleeding oesophageal varies and therefore no delay ourred between the start of the bleeding and the first sample. learly, ILE dereases to approximately 3% of initial values within 12 hours. This illustrates that the initial derease in ILE after upper gastrointestinal bleeding is probably not seen in the other patients beause of the delay between the start of the bleeding and the first sample. The sum of VAL and LEU (branhed hain amino aids; Fig 4) showed a pattern similar to that seen in pepti uler patients; they were inreased after Amino aid, ammonia, and urea onentrations after upper gastrointestinal haemorrhage in pepti uler patients Time ontrol -12 h 12-24 h 24-48 h 48-72 h 72-96 h 96-24 h UREA 5 () 19 (2) 17 (3) 15 (3) 11 (3)11 1 (3)11 9 (3) AMM 25 (1) 26 (5) 24 (3) 2 (3)11 19 (4)11 17 (2)11 21 (3) GLU 5 (2) 68 (2) 47 (1) 46 (8) 42 (8) 47 (9) 53 (11) ASN 47 (1) 52 (6) 52 (4) 46 (4) 54 (6) 54 (4) 52 (8) SER 97 (2) 95 (17) 93 (1) 86 (12) 19 (17) 18 (17) 17 (26) GLN 664 (9) 584 (48) 613 (53) 528 (44)11 572 (49) 56 (71) 541 (75) GLY 23 (8) 169 (29) 167 (22) 149 (15) 187 (18) 2 (16) 22 (29) THR 125 (3) 128 (25) 15 (12) 8 (11) 113 (13) 118 (1) 12 (22) HIS 82 (1) 87 (16) 73 (8) 55 (6) 55 (6) 53 (5) 52 (8)*11 IT 35 (1) 47 (7) 38 (5) 27 (3) 27 (4) 29 (6) 32 (8) ALA 348 (11) 336 (65) 264 (46) 182 (13) 198 (26) 245 (28) 246 (41) TAU 41 (1) 62 (11)11 53 (6) 48 (7) 48 (9) 51 (1) 61 (14) ARG 76 (2) 58 (6) 57 (5)11 53 (6)11 69 (8) 65 (9) 69 (15) aaba 19 (1) 35 (8) 32 (6)11 25 (3)5 23 (3) 2 (4) 16 (3) TYR 59 (1) 85 (17) 69 (1) 51 (1)11 61 (11) 6 (16) 47 (1)11 VAL 233 (4) 524 (88) 352 (44)11 225 (33) 216 (32) 185 (32) 178 (15)1 MET 23 () 18 (2)11 19 (2) 19 (2)11 24 (3) 23 (3) 22 (3) ILE 63 (1) 35 (6) 43 (6)11 48 (5)11 58 (9) 54 (9) 64 (8)* PHE 59 (1) 17 (1) 9 (9) 62 (4) 61 (5) 56 (4) 55 (2): TRP 52 (1) 31 (8) 26 (3) 22 (3) 25 (5) 24 (4)11 22 (4) LEU 129 (3) 223 (49) 155 (18) 1 (12)11 111 (18) 82 (1)11 92 (8)tlI LYS 157 (3) 176 (22) 158 (16) 134 (12) 162 (16) 151 (12) 132 (32) SUMAA 2586 (31) 2927 (319) 254 (166) 1992 (116) 2216 (161)11 2183 (15) 2163 (25)* VAL+LEU 362 (7) 747 (135) 56 (61)11 325 (45) 327 (48) 267 (4) 27 (22):1 AAA 117 (2) 192 (26) 159 (16)11 113 (11) 122 (13) 115 (18) 12 (2)t SUM-ILE 2523 (3) 2891 (32) 2461 (167) 1944 (116) 2158 (155)11 2129 (144)11 298 (242)* Data are presented as means (SEM) in,um alulated as desribed under Statistis, exept urea (mm). ontrol values, obtained in preoperative eletive surgial patients (n= 16), are given in the first olumn. SUMAA: sum of the individual amino aids measured; AAA: aromati amino aids (PHE+TYR). For pratial reasons, only some time points are shown in the Table. Statistis: one way analysis of variane proedure for time effets: *p<.5, tp< 1, tp< 1; Mann-Whitney U proedure versus ontrol values Sp<-5, 11p<-1, p<-1. Gut: first published as 1.1136/gut.39.1.13 on 1 July 1996. Downloaded from http://gut.bmj.om/ on 15 January 219 by guest. Proteted by opyright.
16 Dejong, Meijerink, van Berlo, Deutz, Soeters ± ). 4Ji.) a,.) 5F 1 r A Subjet L L ~~~~Subjet L 15 r 1H 95 5 5 25 1L liver funtion, although it is well known that branhed hain amino aid metabolism is hanged and plasma branhed hain amino aid onentrations are dereased in the seond group.1 In reent years, we have shown that within six hours after intragastri blood administration in Subjet K pigs, as well as in healthy volunteers, a more than 5% derease in plasma ILE values ourred, while the onentration of the other branhed hain amino aids LEU and VAL and - most other amino aids reahed extremely high values.3-5 Simultaneously, arterial urea and ammonia values inreased.3 Isonitrogenous ontrol meals led to a less pronouned inrease of most amino aids, inluding ILE, and a signifiantly less pronouned inrease in arterial Subjet M ammonia and urea. Similar findings were also made in rats with hroni liver insuffiieny.1' o o 2 2 The hanges in plasma amino aids, ammonia, 5 1 15 2 25 5 1 15io 2 25 and urea seen in this study are remarkably similar to those seen in previous studies in our Figure 3: ILE onentrations after major, endosopially verified, upper gasttrointestinal department in experimental animals and in haemorrhage in four individual irrhoti patients (mean age 43 years (rangxe33-5)). The healthy volunteers. This seems to point to a start of bleeding, assessed by areful history taking, is taken as t= hours. ± '._ X 5 ) ) o 2 2 = 1 5 1 ) E universal pathophysiologial mehanism in various speies, whih probably is as follows. gastrointestinal haemorrhage an( d gradually Blood protein (mainly haemoglobin) is dereased to stable values. Ammonia (not almost ompletely devoid of ILE.3 As ILE is an shown) was inreased in all irrh( )ti patients essential amino aid, the absene of ILE after upper gastrointestinal haem4orrhage and implies that blood protein has a low biologial normalised after ammonia lowerinig treatment value.6 Thus, gastrointestinal haemorrhage or (latulose and neomyin in subje t J,K,L, and M, plus dopamine9 in patient K). blood ingestion presents the organism with a protein load that annot be utilised for protein syntheti purposes beause an essential amino aid is laking. This may explain why most of Disussion the amino aids show a onsiderable and more This study shows that ILE dere ases by 6- pronouned rise after upper gastrointestinal 7 /O within six hours after gas;trointestinal haemorrhage than after an isonitrogenous haemorrhage, onfirming previou[s studies in protein load of normal omposition.34 As a our department on the effets of 1blood inges- onsequene, the onstituting amino aids will tion in pigs and healthy volunte:ers.3-5 This be metabolised or degraded, or both, resulting ILE derease ours in patients with intat in ammonia and urea synthesis. This means liver funtion and in patients wiith impaired that the low biologial value of blood protein may explain its ammoniageni and ureageni potential. In aordane with this, supplementing ILE simultaneously with an intra- 5 n Subjet J Subjet K 5 H-_ almost fully be aounted for by diminished U. intestinal ammonia and amino aid release.4 1 o. 5 1 15 2 25 5 1 1E 5 2 25 From these studies it was onluded that intravenous ILE supplementation improved the biologial value of blood protein, seemed to Figure 4: The onentration of the sum of VAL and LEU after major, endo. isopially improve amino aid utilisation, espeially in verified, upper gastrointestinal haemorrhage in four individual irrhoti pati 43 years (range 33-5)). The start of bleeding, assessed by areful history t kng,s(metaaken the small bowel wall, and diminished intestinal as t= hours. ammonia release.4 gastri blood protein load should improve the biologial value of blood protein and redue its ammoniageni and ureageni potential. In full agreement with this, we previously showed that the rise in arterial ammonia and urea values ould be blunted in pigs by ILE infusion after blood administration intrari gastrially.4 In these experiments,4 the higher arterial onentrations of ammonia and several amino aids after intragastri blood adminis- Subjet M tration than after a ontrol meal3 ould be largely aounted for by inreased intestinal ammonia and amino aid release into the irulation. The smaller rise in arterial ammonia and amino aids (exluding ILE) in pigs reeiving ILE infusion after a blood meal ould Gut: first published as 1.1136/gut.39.1.13 on 1 July 1996. Downloaded from http://gut.bmj.om/ on 15 January 219 by guest. Proteted by opyright.
Plasma isoleuine after gastrointestinal bleeding 17 The ause of the 6-7% plasma ILE derease is speulative. It seems self evident that after a protein meal defiient in ILE, this amino aid will not inrease in plasma. However, this does not explain a derease. In blood protein, methionine is similarly defiient,3 but only dereases by 25% after gastrointestinal haemorrhage in uler patients, suggesting that additional fators play a part in the observed hanges. Branhed hain amino aid antagonism12-15 may play a part in the ILE derease.4 Two mehanisms may be important in this respet. Firstly, the inreased onentrations of VAL and LEU may ompete with the dereased TIL values for transport aross the ell membrane. Thus, the ILE derease might be even more profound intraellularly. Evidene to support this has been found reently in rats with hroni liver failure, where we found a derease in hepati tissue ILE onentrations after intragastri blood administration.11 Seondly, inreased LEU values stimulate the ativity of the branhed hain 2-oxoaid dehydrogenase omplex, the rate limiting step in branhed hain amino aid breakdown. Thus, even at these diminished ILE onentrations ILE oxidation might be inreased. 14 The proposed mehanism (diminished protein synthesis by ILE defiieny) is onsistent with findings by others onerning impaired protein synthesis, DNA repliation, ell multipliation, and ell funtion in an ILE defiient state.1618 Impaired protein synthesis ould shift the balane between synthesis and degradation towards degradation, induing a net ataboli state. This an be expeted to primarily affet rapidly dividing ells, short half life proteins like some lotting fators, and aute phase proteins, as well as proteins rih in ILE. In irrhoti patients, hyperammonaemia after upper gastrointestinal bleeding may additionally impair lymphoyte funtion.19 The finding in pigs, that part of the adverse metaboli hanges indued by blood protein administration ould be reversed by intravenous ILE administration ould point to possibilities for metaboli support in upper gastrointestinal bleeding patients. This ould improve amino aid utilisation in protein syntheti proesses and thereby lower intestinal ammonia generation.4 Diminished intestinal ammonia generation an be expeted to lead to redued urea synthesis and hene, ILE supplementation ould redue uraemia after gastrointestinal haemorrhage. In onlusion, these data show that after upper gastrointestinal haemorrhage in otherwise healthy humans, the plasma onentration of the amino aid ILE dereases preipitously, whereas monst other amino aids inrease. This onfinns previous studies in experimental animals and healthy volunteers. Thus, we hypothesise that, in analogy with the experimental animal situation, intravenous ILE administration during or after upper gastrointestinal bleeding ould be used to improve nitrogen eonomy in bleeding uler patients. The limited data from patients with liver irrhosis presented in this study seem to be onsistent with this hypothesis. Future studies will fous on supplementing ILE during upper gastrointestinal haemorrhage in humans with and without impaired liver funtion. The authors wish to thank Mr H M H van Eijk and Mrs M A H van der Heijden for their skilled amino aid determinations. 1 Bessman AN, Mirik GS, Hawkins R. Blood ammonia levels following the ingestion of asein and whole blood. J lin Invest 1958; 37: 99-9. 2 Rudman D, DiFulo TJ, Galambos JT, Smith RB, Salam AA, Warren WD. Maximal rates of exretion and synthesis of urea in normal and irrhoti subjets. J lin Invest 1973; 52: 2241-9. 3 van Berlo LH, van de Boogaard AEJM, van der Heijden MAH, van Eijk HMH, Janssen MA, Bost MF, et al. Is inreased ammonia liberation after bleeding in the digestive trat the onsequene of omplete absene of isoleuine in hemoglobin? A study in pigs. Hepatology 1989; 1: 315-23. 4 Deutz NEP, Reijven PLM, Bost MF, van Berlo LH, Soeters PB. Modifiation of the effets ofblood on amino aid metabolism by intravenous isoleuine. Gastroenterology 1991; 11: 1613-2. 5 van Berlo LH, Dejong H, Meijerink WJHJ, Fik TE, von Meyenfeldt MF, Deutz NEP, et al. A derease in plasma isoleuine levels after simulated upper digestive trat bleeding in man [Abstrat]. lin Nutr 1991; 1 supple 2: 56. 6 Newshoime EA, Leeh AR. Biohemistry for the medial sienes. New York: John Wiley, 1983. 7 van Eijk HMH, van der Heijden MAH, van Berlo LH, Soeters PB. Fully automated liquid-hromatographi determination of amino aids. lin hem 1988; 34: 251-3. 8 van Eijk HMH, Dejong H, Deutz NEP, Soeters PB. Influene of storage onditions on normal amino aid onentrations. lin Nutr 1994; 13: 374-8. 9 Zieve L, Doizaki WM, Derr RF. Reversal of ammonia oma in rats by L-DOPA: a peripheral effet. Gut 1979; 2: 28-32. 1 Weber FL, Bagby BS, Liate L, Kelsen SG. Effets of branhed hain amino aids on nitrogen metabolism in patients with irrhosis. Hepatology 199; 11: 942-5. 11 Olde Damink SWM, Dejong H, Deutz NEP, Soeters PB. Intestinal and hepati tissue isoleuine levels after simulated upper gastrointestinal hemorrhage in two models of hroni liver insuffiieny in rats. [Abstrat]. lin Nutr 1993; 12 (suppl 2): 43. 12 Rogers QR, Spolter PD, Harper AE. Effet of leuineisoleuine antagonism on plasma amino aid pattern of rats. Arh Biohem Biophys 1962; 97: 497-54. 13 Oestemer GA, Hanson LE, Meade RJ. Leuine-isoleuine interrelationship in the young pig. Jf Anim Si 1973; 36: 674-8. 14 Aftring RP, Blok KP, Buse MG. Leuine and isoleuine ativate skeletal musle branhed-hain a-keto aid dehydrogenase in vivo. Am J Physiol 1986; 25: E599-64. 15 Swendseid ME, Villalobos J, Figueroa WS, Drenik EJ. The effets of test doses of leuine, isoleuine or valine on plasma amino aid levels. The unique effet of leuine. Am Jlin Nutr 1965; 17: 317-21. 16 Bergstrom J, First P, Vinnars E. Effet of a test meal, without and with protein, on musle and plasma free amino aids. lin Si 199; 79: 331-7. 17 Kohn A. Differential effets of isoleuine deprivation on ell motility, membrane transport and DNA synthesis in NILS hamster ells. Exp ell Res 1975; 94: 15-22. 18 Tobey RA, Ley KD. Isoleuine-mediated regulation of genome repliation in various mammalian ells lines. aner Res 1971; 31: 46-51. 19 Zimber A, Visek WJ. Effet of urease injetions on DNA synthesis in mie. Am J Physiol 1972; 223: 14-8. Gut: first published as 1.1136/gut.39.1.13 on 1 July 1996. Downloaded from http://gut.bmj.om/ on 15 January 219 by guest. Proteted by opyright.