ABNORMAL HEPATIC TRANSPORT OF INDOCYANINE GREEN IN GILBERT'S SYNDROME

Transcription

1 GASTROENTEROLOGY 70:38&-391, by The Williams & Wilkins Co. Vol. 70, No.3 Printed in U.S.A. ABNORMAL HEPATIC TRANSPORT OF INDOCYANINE GREEN IN GILBERT'S SYNDROME JAMES F. MARTIN, M.D., JOHN M. VIERLING, M.D., ALLAN W. WOLKOFF, M.D., BRUCE F. SCHARSCHMIDT, M.D., JOHN VERGALLA, A.B., JEANNE G. WAGGONER, B.A., AND PAUL D. BERK, M.D. Section on Diseases of the Liver, Digestive Diseases Branch, National Institute of Arthritis, Metabolism and Digestive Diseases, National Institutes of Health, Bethesda, Maryland The plasma fractional disappearance rate of indocyanine green (k IC (;' min- I) and the absolute hepatic ICG removal rate (V IC (;, mg per kg per min) ere determined in 26 patients ith Gilbert's syndrome (GS) and 19 normal control volunteers folloing intravenous administration of doses of 0.5, 2.0, 3.5, and 5.0 mg per kg of dye. The diagnosis of GS as based on studies of radiobilirubin kinetics in all cases and liver biopsy in 22 of 26 cases. The patients ere further classified into 3 subgroups, based on patterns of sulfobromophthalein (BSP) kinetics, as follos: GS 1(15 patients) had normal BSP metabolism; GS II (5 patients) had a defect in BSP metabolism beyond the stage of initial hepatic uptake; and GS III (6 patients) had a defect in the initial hepatic uptake ofbsp (Gastroenterology 63: ,1972). Both k 1c (; and V 1CG ere significantly reduced, compared to normal controls, in the GS III group ith defective BSP uptake, but did not differ significantly from normal in the GS I and GS II groups. Michaelis Menten analysis of the data indicated that V MAX for ICG uptake in the GS III group (1.2 ± 0.6 mg per min per kg) as significantly reduced compared to the previously established normal value of 3.6 ± 0.6 mg per min per kg; (P < 0.01). For the total population of 26 patients ith Gilbert's syndrome, there as a highly significant correlation (r = 0.77, P < 0.01) beteen k 1c (: and A 2B I~ P the ' fractional hepatic uptake rate for BSP. These studies confirm previous ork indicating that patients ith Gilbert's syndrome constitute a heterogeneous population ith regard to defects in hepatic organic anion transport, some of the defects not being attributable to glucuronyl transferase deficiency. Future studies of Gilbert's syndrome must take into account the existence of these subgroups, since they may have different underlying pathogenetic mechanisms. Gilbert's syndrome is a common and frequently familial disorder characterized by mild chronic unconjugated hyperbilirubinemia, not attributable to hemolysis, occurring in the absence of structural liver disease. la, 2 Patients ith this condition characteristically have a reduction in hepatic bilirubin clearance (C BR) hich permits it to be diagnosed even in the presence of concomitant hemolysis. 3 The reduction in C BR has been idely,4-6 although not universally,7 attributed to a partial deficiency of the conjugating enzyme bilirubinuridine diphosphate glucuronyl transferase (UDPGT). Received June 9, Accepted October 3, Portions of these data ere presented at the 31st Annual Meeting, American Federation for Clinical Research, Atlantic City, Ne Jersey, May 5, 1974, and have been published in abstract form.' Address reprint requests to: Dr. Paul D. Berk, Bldg. 10, Rm. 4D-52, Digestive Diseases Branch, National Institute of Arthritis, Metabolism and Digestive Diseases, National Institutes of Health, Bethesda, Maryland The authors are indebted to the nursing staff of the Liver Unit for excellent care of the patients in this paper and assistance ith the studies described, to Ms. Phyllis Melton for cheerful and accurate preparation of the manuscript, and to Dr. E. Anthony Jones for his critical revie of the manuscript. 385 Although the hepatic transport of organic anions other than bilirubin has generally been thought to be normal in Gilbert's syndrome, it as recently reported that a minority of otherise typical patients had abnormalities of sulfobromophthalein (BSP) metabolism. 8 In the present study, plasma disappearance rates of graded doses of indocyanine green (leg), another organic anion cleared from the circulation by the liver, ere determined in patients ith Gilbert's syndrome having both normal and abnormal BSP metabolism, and in normal volunteer control subjects. The results confirm both the heterogeneity of the patient population ith Gilbert's syndrome and the existence of a subgroup ith a defect in the hepatic uptake of organic anions hich is not attributable to glucuronyl transferase deficiency.8 Materials and Methods Patients Beteen July 1, 1967, and February 1, 1975,55 patients ith Gilbert's syndrome ere evaluated at the Clinical Center of the National Institutes of Health. Tenty-six of these patients, including 21 males and 5 females, ranging in age from 9 to 66 years (mean = 29 years), ere available to participate in the

2 386 MARTIN ET AL. Vol. 70, No. 3 studies described belo. Ten of these 26 individuals, including 6 of the 11 ith abnormal BSP kinetics, ere included in a previous report describing abnormalities ofbsp metabolism in Gilbert's syndrome. 8 In each of the 26 patients, the diagnosis as based initially on the clinical finding of chronic or recurrent unconjugated hyperbilirubinemia in the presence of repeatedly normal values for serum glutamic oxaloacetic transaminase (SGOT) and serum glutamic pyruvic transaminase (SGPT), alkaline phosphatase, and serum protein electrophoresis, and as confirmed by the characteristic finding of reduced plasma clearance of radiobilirubin.' Percutaneous liver biopsies ere obtained or available for revie from 22 patients, including 9 of the 11 ith abnormal BSP metabolism (vide infra). Apart from prominence of lipofuscin pigment, hich as particularly striking in t he centrilobular regions in carbolfuchsin-stained sections, 9 hepatic histology as normal in all cases. Eight of these 26 patients had concomitant hemolytic disease, documented by a shortened SlCr red blood cell half-life. 3, 8, 10, 11 In three instances, hemolysis as due to hereditary spherocytosis and in one to glucose-6-phosphate dehydrogenase deficiency. No specific etiology as discernible in the other four cases, despite extensive investigation. Sixty normal volunteers beteen the ages of 18 and 29 served as controls for the various studies described belo: 19 (9 men, 10 omen) for the leg studies, 17 ( 9 men, 8 omen) for the BSP studies, and 24 (11 men, 13 omen) for the radiobilirubin studies. Volunteers met strict, previously described criteria for "normality."8,'2 Studies ere performed ith the fully informed consent o f all participants. The protocols employed ere approved by the Radiation Safety and Clinical Research Committees and the Medical Board of the National Institutes of Health. None of the participants as taking any medication at the time of the studies described belo. Experimental Studies L Plasma radiobilirubin disappearance studies. Plasma radiobilirubin disappearance curves of 30-hr duration ere obtained as previously described 12 using either [UC ]bilirubin" or [ 3H]bilirubin. U The curves ere analyzed by means of the Simulation, Analysis and Modeling (SAAM) program of Berman and Weiss'S on a Univac 1108 digital computer in order to determine bilirubin clearance (C HH : ml per min per kg) and da ily plasma bilirubin turnover (BRT: mg per kg per day), the calculation of hich has been shon to be independent of any particular model of bilirubin metabolism, 12 The plasma unconjugated bilirubin concentration (BR: mg per 100 m!) used in these calculations as the mean of at least 9 individual measurements, performed by the method of Weber and Schalm, '6 during the course of the radio bilirubin study. In addition to these model-independent calculations, the curves ere fitted by computer to a previously described 12 and validated l7 three-compartment model of bilirubin metabolism in order to calculate such variables as A21> the fractional hepatic uptake rate for unconjugated bilirubin (min-1) and A02, the fractional conjugation rate o f the intrahepatic unconjugated bilirubin pool (min- I). 2. Plasma BSP disappearance studies, Plasma BSP disappearance curves ere obtained for a period of 60 min, as previously described,8 f olloing the intravenous injection of the standard dose of 5.0 mg per kg. Plasma BSP concentration as measured by the method of G aebler,'8 and 45-min BSP retention as determined from the plasma concentration of the 45-min sample using the conventional assumption of a value a t zero time of 10.0 mg per 100 m!. As reported previously," the BSP curves ere also analyzed by computer to calculate t he model-independent variables he (the fraction of the plasma BSP pool irreversibly extracted from the circulation (min - ')) and BSP clearance (C BS!', ml per min per kg), as ell as a, b, and h, the fractional transfer rates of a to-compartment model of BSP metabolism. Under a long-used notation, 19, 20 a represents the model-derived fractional transfer rate of BSP from plasma to liver (min-i), and band h are the corresponding fractional transfer rates from liver back to plasma and to bile, respectively, 3. Plasma leg disappearance studies. Blood samples ere obtained 3.5, 5, 6.5, 8,10, 12, 15, 20, 25, 30and, usually, 60 min after the rapid intravenous injection oficg at doses of 0.5, 2. 0, 3.5, and 5.0 mg per kg. In general, to studies ere performed on each of to separate days, the loer dose (either 0.5 o r 2.0 mg per kg) for the day being administered first. At least 2 hr elapsed beteen the injection of the to doses. Preliminary studies established that, hen conducted in this manner, prior administration of the first dose did not influence the resul ts obtained ith the second. In a fe instances, only a single study as performed on each of 4 days; but in no case ere more than to studies performed ithin any 24-hr period. Plasma ICG concentration as determined from the optical density at 805 nm,21 employing a standard curve prepared for each study by adding graded amounts of the injected stock solution to the subject's on plasma. The plasma ICG concentration as plotted semilogarithmically versus time. T he data points beteen 5 and 12 min ere linear, and the slope of a least squares fit to this portion of the curve as defined a s the plasma ICG fractional disappearance rate (h[(.", min- I). The actual plasma ICG removal rate VI C" (mg per min per kg) as calculated from the relationship V ice: = k x dose (mg per kg). Because ICG has little extravascular distribution or extrahepatic metabolism 21, 22 its plasma removal rate, V, closely approximates its rate of hepatic uptake. 23 For each population of patients o r normal volunteers described belo, a plot o f V versus dose took the form of a rectangular hyperbola analyzable in terms of Michaelis Menten kinetics. Accordingly, for each population, these data ere fitted by computer directly to the Michaelis-Menten equation in order to calculate V M.", (mg per min per kg), the maximal rate of hepatic ICG uptake, and KM (mg per kg), the dose at hich half-maximal uptake as achieved. Except for computational differences based on the computer and program employed, this procedure as essentially that of Paumgartner et a!. 23 Results As previously reported, the patients ith Gilbert's syndrome exhibited 3 patterns of BSP metabolism. In 15, plasma BSP disappearance curves ere entirely normal, as defined by normal values (ithin 2.0 so of the normal mean) for 45-min BSP retention, ke and CHS I" These patients ere designated as Gilbert's syndrome (GS) I. Patients ith abnormal BSP transport as manifested by abnormal values for any of the three variables listed above ere classified as GS II if hepatic BSP uptake (model variable a) as normal (5 individuals) or GS III if hepatic BSP uptake as reduced (6 individuals). These classification criteria, hich differ slightly from those previously reported, permit improved recognition of patients ho have an abnormal BSP disappearance curve but ho, nevertheless, have normal 45-min BSP retention.

3 March 1976 ABNORMAL HEPA TIC TRANSPORT 387 Forty-five minute BSP retention values for each of the 3 subtypes of Gilbert's syndrome are presented in figure 1, and the variables determined by kinetic analyses of BSP disappearance curves are presented in table 1. By z o f Z f a: CL 10 (f) m f- :J Z 2 LD " 5 I., O L ~ ~ ~ ~ L GS I GS II Gsm FIG. 1. Forty-five minute sulfobromophthalein (BSP) retention in patients ith Gilbert's syndrome grouped according to pattern of BSP metabolism (see text for definition of groups). Dashed line represents upper limit of normal range. Group No.of individuals definition, all aspects of BSP transport ere normal in group GS I. Forty-five minute BSP retention as abnormal in 3 of 5 patients in group GS II and 5 of 6 patients ith GS III. Both he and C H ere ~ P significantly reduced (Student's t-test, P < 0.01) in GS II and GS III hen compared to either normal volunteers or GS I. Although these variables tended to be loer in GS III than GS II, differences in he (P > 0.1) and C H ~ (P> I ' 0.2) beteen the latter subgroups ere not statistically significant. In accordance ith the definition of the subgroups, hepatic BSP uptake (model variable a) as normal in GS II, in hom the reduction in C H ~ as P the result of an abnormality at a later stage of BSP transport. 8 In contrast, hepatic BSP uptake as reduced, but other model-dependent variables ere normal, in GS III. The plasma leg fractional disappearance rates at various dose levels are presented in table 2. There ere no significant differences beteen normal volunteers and patients ith either GS I or GS II at any of the doses employed. In contrast, h rc (; in the GS III patients as significantly reduced (P < 0.01) at all of the doses employed hen compared to either normal volunteers or patients ith GS I. The plasela leg removal rate, V 1CG, equivalent to the hepatic uptake rate, is plotted as a function of dose in figure 2. Hepatic uptake of leg did not differ significantly from normal control subjects in either GS I, in hich BSP transport as normal, or GS II, in hich the abnormality in BSP transport appeared to be at a stage beyond hepatic uptake. Over the dose range of 0.5 to 5.0 mg per kg, neither normal subjects nor patients ith GS I or GS II shoed clear-cut saturation of leg uptake ith TABLE 1. Parameters of BSP metabolism in patients ith Gilbert's syndrome" Model-independent parameters CBS/' k, (ESP)" a' mlpermin per kg min- 1 min- 1 Normal volunteers ± ± ± ( ) ( ) ( ) Parameters of BSP' model b ' h' min- 1 min ± 0.002h 0.029± ( ) ( ) GS'I ± ± ± ( ) ( ) ( ) GS II ± 0.6' ± 0.015' ± ( ) ( ) ( ) GS III ± 004' ± 0.017' ± h., ( ) ( ) ( ) ± J ± ( ) ( ) ± ± ( ) ( ) ± ± ( ) ( ) " All values mean ± SD. Numbers in parentheses represent observed ranges. b BSP, sulfobromophthalein. C CH,p, BSP clearance. d he (BSP), fraction of plasma BSP pool irreversibly removed per minute. e a, fraction of plasma BSP pool entering liver per minute. I b, fraction of hepatic BSP pool refluxing to plasma per minute. h, fraction of hepatic BSP pool transferred into bile per minute. h Compared ith Gilbert's syndrome (GS) II; P < 0.01 (Student's t-test)., GS, Gilbert's syndrome. J Compared ith GS II; P < 0.02 (Student's t-test). 'Compared ith normal volunteer subjects and GS I; P < 0.01 (Student's t-test).

4 388 MARTIN ET AL. Vol. 70, No.3 TABLE 2. Plasma indocyanine green fractional disappearance rate (hie,,) for patients ith Gilbert's syndrome (min -1)a ---- Administered dose (mg/kg) Normal volunteers, n ~ ± ± ± ± GS 1. n ~ ± ± ± ± GS II, n ~ ± ± ± ± 0.031" GS III, n ~ ± 0.041" ± O.017 c ± c ± 0.031" a All values mean ± SD. "Compared to GS III; P < 0.01 (Student's t-test). C Compared to normal volunteer subjects and GS I; P < 0.01 (Student's t-test). 1.0 GS I 0.3 OJ "" c E OJ E 0.8 ~ 06 2 f «a: - ~ «> a: c:j u «:? (f) ~ 04 «...J (L 0.2 I Mean±SEM Gsn GSm 0 ~ - - ~ L L J ~ ~ leg DOSE Img/kg) FIG. 2. Plasma indocyanine green (lcg) removal rate (V'c,,) as a function of administered dose in normal control subjects and patients ith Gilbert's syndrome (GS), grouped according to sulfobromophthalein metabolism. GS III differs significantly from normal control subjects and GS I at all dose levels (P < 0.01). increasing dose. In contrast, a statistically significant reduction of leg uptake at all doses, and obvious saturation of uptake ith increasing doses as manifested by GS III patients. Thus, in the various Gilbert's syndrome subgroups, the hepatic uptake of leg appears to parallel the hepatic uptake of BSP, as represented by variable a of the compartmental model of BSP metabolism. In figure 3 the fractional hepatic uptake rate of BSP is plotted against kic(j (5.0 mg per kg dose) for all 26 patients ith Gilbert's syndrome. These to variables ere significantly correlated (r = 0.77, P < 0.01). Values for the Michaelis-Menten analysis of leg t E./ 0/ "'d ",/ "" t>',. 0 '; c 0.1 GS [ '" o GS II ~ '" GSm O L L ~ - Ql ~ ~ FIG. 3. Comparison beteen fractional hepatic uptake rate of sulfobromophthalein (BSP) (a, "21 BS p) and indocyanine green (ICG) (h,cd in all 26 patients ith Gilbert's syndrome (GS). uptake are presented in table 3. Over the restricted range of doses employed, there as only a small amount of curvature to the plots of V versus dose in the normal control subjects and in patients ith GS I and GS II (fig. 2). Hence, there is considerable statistical imprecision, represented by large standard deviations, for the value of V MAX and k m in these groups. The mean values for V MAX are, nevertheless, similar to the value for normal men of 3.6 ± 0.6 mg per min per kg reported by Paumgartner et al. in a study employing higher doses up to 10.0 mg per kg.23 In contrast, both V MAX and Km are ell defined by the current dosage range in GS III. The value for V MAX, 1.2 ± 0.6 mg per kg per min, is significantly reduced compared to the normal value reported by Paumgartner et al. 23 (P < 0.01) and confirms the presence of a defect in hepatic organic anion uptake in this subgroup. Because of the idely accepted belief that bilirubin, BSP, and leg share a common mechanism for hepatic uptake,2 an attempt as made to relate abnormalities of BSP and leg kinetics to those of bilirubin metabolism. As indicated in table 4, plasma unconjugated bilirubin concentration as significantly increased and C BR significantly reduced, compared to normal, in all three

5 March 1976 ABNORMAL HEPA TIC TRANSPORT 389 types of Gilbert's syndrome. Furthermore, BRT as elevated, reflecting the presence of patients ith hemolysis, in all groups. With regard to model-dependent parameters of bilirubin metabolism, the values for.\02 (conjugation) and.\21 (uptake) ere also significantly reduced in all three types of Gilbert's syndrome. Although.\21 as loest in GS III, there ere no statistically significant differences in any aspect of bilirubin metabolism among the three Gilbert's syndrome subgroups. Discussion The present data confirm the conclusions of an earlier study from this laboratory,8 namely, that (1) Gilbert's syndrome represents a heterogeneous population ith regard to defects in hepatic organic anion transport; (2) although the defect may be limited to bilirubin transport in most cases, there are to subpopulations of Gilbert's syndrome ho have abnormalities in the hepatic transport of organic anions other than bilirubin; and (3) of these subgroups, one has a defect in the hepatic uptake of organic anions, and another a defect at a later stage of their transhepatic transport. The nature of this latter defect remains obscure, and the present study sheds no light on this issue. Because of the agreement beteen the findings obtained ith ICG and BSP, and the high degree of correlation beteen the fractional rate of hepatic BSP TABLE 3. Michaelis-Menten analysis of hepatic uptake of ICG in patients ith Gilbert's syndrome a V mu, Km' mg/kg/ mm mg/kg Normal volunteer, n ~ ± ± 24.2 GS I, n ~ ± ± 51.0 GS II, n ~ ± ± 12.4 GS III, n ~ ± ± 3.7 a All values, mean ± SD. b V maxo maximum hepatic ICG uptake,rate., Km, dose producing one-half the maximum uptake rate. uptake (a) and k[cg, the study also supports the concept that BSP anci ICG share a common hepatic uptake mechanism. It also provides further support for the validity and usefulness, in man, of the to-compartment model of BSP metabolism employed in the analysis of the data. This model had previously been most extensively evaluated in the dog. 19 Several questions are raised by the current study. With regard to the underlying defect(s) in Gilbert's syndrome, a consistent reduction of UDPGT has been reported by several groups4-6, 24 and confirmed in an unreported series from this department. It has been suggested, therefore, that Gilbert's syndrome is due to a partial deficiency of UDPGT, and that it differs from the congenital nonhemolytic jaundice syndromes 25 only in the severity of the enzymatic defect. 26 The defects in ICG and BSP transport seen in GS II and GS III do not appear explainable on the basis of a reduced UDPGT, because ICG is excreted into the bile ithout prior conjugation 21, 22, 27 and BSP is conjugated, in part, by a soluble enzyme 28 or enzymes 29 to glutathione. Several other lines of evidence had previously suggested that partial deficiency of UDPGT as inadequate to explain completely the pathophysiology of Gilbert's syndrome. In particular, compartmental analysis of radiobilirubin disappearance curves (table 4) suggested the presence of a defect in hepatic bilirubin uptake in all three types of Gilbert's syndrome. 3, 8 Hoever, hepatic bilirubin uptake as virtually normal in an adult patient ith type 1 congenital nonhemolytic jaundice (Crigler Najjar syndrome)30 and bilirubin rapidly enters the liver of jaundiced Gunn rats. 31 In both instances, UDPGT is completely lacking. Hence, the apparent bilirubin uptake defect in Gilbert's syndrome ould not appear to be attributable to a partial deficiency of this enzyme. Furthermore, administration of drugs such as phenobarbital to patients ith Gilbert's syndrome corrects the abnormality in bilirubin metabolism 26, 32 ithout significantly altering UDPGT. 24,33 Although the hypothesis that defective UDPGT is the TABLE 4. Parameters of bilirubin metabolism in patients ith Gilbert's syndrome (GS)a HR' CH1./ BRT' A21 e Au I >"02 11 k. (BR)" mg/looml ml/min/kg mg/kg/day I Normal volunteers, 0.44 ± ± ± 0.68 n ~ 24 GS I, n ~ ± 0.79' 0.21 ± 0.07' 5.38 ± 1.19' GS II, n ~ ± 0.64' 0.21 ± 0.06' 4.58 ± 0.58; GS III, n ~ ± 0.35' 0.20 ± 0.04' 5.16 ± 1.06' min ± ± ' ± ' ± ' min ± ± ± ± min- 1 min ± ± ± ' ± ' ± ' ± ' ± ' ± ' a All values mean ± SD. b BR, plasma unconjugated bilirubin concentration. C C RR, hepatic bilirubin clearance. d BRT, plasma bilirubin turnover. e A'b fractional transfer rate for hepatic uptake of bilirubin. f A I2, fractional transfer rate for ref1ux of bilirubin from liver to plasma. A o " fractional transfer rate for irreversible removal of bilirubin from the system ("conjugation"). h ke (BR), fraction of bilirubin plasma pool irreversibly cleared per minute. 'Compared ith normal volunteer subjects; P < 0.01 (Student's t-test). ; Compared ith normal volunteer subjects; P < 0.02 (Student's t-test).

6 390 MARTIN ET AL. Vol. 70, No.3 basis for Gilbert's syndrome does not appear to explain the data outlined above, information important to the formulation of a ne hypothesis is currently lacking. In particular, it ould be important to kno (1) hether the reduced values for UDPGT reported in Gilbert's syndrome as a hole are, in fact, found in patients ith GS II and GS III; (2) hether further studies could demonstrate any differences in the hepatic uptake of bilirubin in GS III patients compared ith those ith GS I and GS II; and (3) the extent to hich the in vitro assay of UDPGT accurately reflects the in vivo activity of the enzyme. Because abnormalities on a BSP test have traditionally excluded patients from the diagnosis of Gilbert's syndrome, no cases of GS II and GS III are likely to have been included in the published studies on UDPGT! In our on series, none of the patients in GS II and GS III had this enzyme assay performed at the time of biopsy because the biopsy as performed either at another institution or before subgroups of Gilbert's syndrome ere recognized. Measurement of UDPGT in GS II and GS III ould be of great interest, but such studies must aait the discovery of ne GS II and GS III patients, as e do not feel justified in repeating a biopsy solely for this purpose. To the extent that bilirubin is believed to share the same hepatic uptake mechanism as BSP and ICG, the data in table 4 are somehat puzzling. Although all three groups of Gilbert's syndrome patients had a reduction in.\21> the fractional hepatic uptake for bilirubin, hen compared to normal controls, there as no difference beteen the GS III group and the others, as might have been expected from the BSP and ICG studies. One possible explanation for this observation may be the fact that.\21 as determined from the injection of a tracer dose of radiobilirubin into subjects ith an endogenous bilirubin concentration averaging less than 2 mg per 100 ml. This is the range of concentration hich ould be produced by injection of a 1 mg per kg (1.7 Jlmol per kg) load of bilirubin. As indicated in figure 2, differences in hepatic ICG uptake among the three Gilbert's syndrome groups ere less obvious at the loer doses of 0.5 to 2.0 mg per kg (0.6 to 2.6 Jlmol per kg) than at doses of 3.5 to 5.0 mg per kg (4.5 to 6.5 Jlmol per kg). In fact, in our previous report, involving a somehat smaller series, and using a dose of 0.5 mg per kg oflcg, differences in k]c(; beteen normal controls and GS III ere not statistically significant.8 It is possible that studies of bilirubin metabolism involving the bolus injection of increasing doses of pigment ould demonstrate a distinction in hepatic bilirubin uptake beteen GS III patients, on the one hand, and GS I and GS II, on the other. The relevance of the UDPGT assay to in vivo conjugating activity has been difficult to study. Although there are documented situations in hich in vitro assays have failed to reflect the biological activity of enzymes in vivo,3' the clear-cut difference in the UDPGT assay beteen Gilbert's syndrome and normal subjects requires explanation. Pending a further evaluation of this problem, any further hypothesis about the mechanisms responsible for Gilbert's syndrome must account for both the reduction in UDPGT activity and defective hepatic uptake of organic anions not attributable to this enzymatic abnormality. A final question concerns the relative incidence of the three subgroups of Gilbert's syndrome. In the National Institutes of Health series, 20 % of cases have had either a GS II or GS III pattern of abnormal BSP metabolism. Because the National Institutes of Health patients are seen by referral, and several of the patients ith abnormal BSP studies ere referred specifically because of this problem, the over-all incidence of these to subgroups is certainly loer than 20% of the total Gilbert's syndrome population. Nevertheless, cases ith abnormal BSP studies are being recognized at other institutions 35 and 1 of 10 consecutive unselected cases of Gilbert's syndrome evaluated by means of BSP kinetics at a large Portuguese hospital as found to be GS II (Carrilho Ribeiro, J. M. Personal communication). The existence of cases of Gilbert's syndrome ith abnormal BSP and ICG transport has certain clinical implications. In patients hose clinical and laboratory findings are typical of Gilbert's syndrome and hose BSP and ICG transport normal, prolonged observation ould appear to be sufficient to establish the diagnosis. In the presence of an abnormal BSP or ICG study, e ould be more inclined to include a liver biopsy in the evaluation to rule out otherise occult liver disease. The identification of several distinct subgroups of otherise typical Gilbert's syndrome also has important implications for the design of future studies of this condition. To the extent that the different subgroups may have differing underlying mechanisms, it is important that future studies be performed in homogeneous populations of appropriately classified patients. REFERENCES 1. Martin JF, Scharschmidt BF, Berk PD: Abnormal indocyanine green kinetics in Gilbert's syndrome (abstr.). Clin Res 22:364a, 1974 lao Foulk WT, Butt HR, Oen CA, et al: Constitutional hepatic dysfunction (Gilbert's disease): its natural history and related syndromes. Medicine 38:25-46, Berk PD, Berlin NI, Hoe RB: Disorders of bilirubin metabolism. In Duncan's Disease of Metabolism, Seventh edition. Edited by PK Bondy, L Rosenberg. Philadelphia, WB Saunders Co, 1974, p Berk PD, Bloomer JR, Hoe RB, et al: Constitutional hepatic dysfunction (Gilbert's syndrome): a ne definition based on kinetic studies ith unconjugated radio-bilirubin. Am J Med 49: , Black M, Billing BH: Hepatic bilirubin UDP glucuronyl transfer ase activity in liver disease a nd Gilbert's syndrome. N Engl,J Med 280: , Metge WR, Oen CA, Foulk WT, et al: Bilirubin glucuronyl transferase activity in liver disease. J Lab Clin Med 64: Frezza M, De Sandre G, Perona G: Hepatic glucuronyl t ransferase activity in congenital and acquired liver disease. Acta Hepato Splenologica 17: , Schmid R, Hammaker L: Glucuronyl transferase formation in

7 March 1976 ABNORMAL HEPA TIC TRANSPORT 391 patients ith constitutional hepatic dysfunction (Gilbert's disease). N Engl J Med 260:131O-1:n4, Berk PD, Blaschke TF, Waggoner JG: Defective BSP clearance in patients ith constitutional hepatic dysfunction (Gilbert's syndrome). Gastroenterology 63: , Barth RF, Grimley PM, Berk PD, et al: Excess lipofuscin accumulation in constitutional hepatic dysfunction. Am Med Assoc Arch Pathol 91:41-47, Sterling K, Gray SJ: Determination of the circulating red cell volume in man by radioactive chromium. J Clin Invest 29: , Read RC: Studies of red cell volume and turnover. Description of a ne "closed" method of red cell volume measurement. N Engl J Med 250: , Berk PD, Hoe RB, Bloomer,JR, et al: Studies of bilirubin kinetics in normal adults. J Clin Invest 48: , Barrett PVD, Mullins FX, Berlin NI: Studies on the biosynthetic production of bilirubin- 14 C: an improved method utilizing deltaaminolevulinic acid-c 14 in dogs. J Lab Clin Med 68: , Hoe RB;"Berk PD, Bloomer JR, et al: Preparation and properties of specifically labeled radiochemically stable 3H-bilirubin.,J Lab Clin Med 75: , Berman M, Weiss MF: User's Manual for SAAM. United States Public Health Service Publication No United States Department of Health, Education and Welfare. Washington, D. C., United States Government Printing Office, Weber AP, Schalm L: Quantitative separation and determination of bilirubin and conjugated bilirubin in human serum. Clin Chim Acta 7: , Berk PD: Total body handling of bilirubin. In Jaundice. Edited by MM Fisher, CA Goresky. Ne York, Plenum Press, Gaebler OH: Determination of bromsulphalein in normal, turbid, hemolyzed, or icteric serums. Am J Clin Pat hoi 15: , Richards TG, Tindall VR, Young A: A modification of the bromosulphthalein liver function test to predict the dye content of the liver and bile. Clin Sci 18: , Barber-Riley, G, Goetzee AE, Richards TG, et al: Tbe transfer of bromsulphthalein from the plasma to the bile in man. Clin Sci 20: , Cherrick GR, Stein SE, Leevy CM, et al: Indocyanine green: Observations on its physical properties, plasma decay, and hepatic extraction. J Clin Invest 39: , Caesar J, Shaldon S, Chiandussi L, et al: The use of indocyanine green in the measurement of hepatic blood flo and as a test of hepatic function. Clin Sci 21:43-57, Paumgartner G, Probst P, Kraines R, et al: Kinetics of indocyanine green removal from the blood. Ann N Y Acad Sci 170: , Felsher BF, Craig JR, Carpio N: Hepatic bilirubin glucuronidation in Gilbert's syndrome. J Lab Clin Med 81: , Arias 1M, Gartner LM, Cohen M, et al: Chronic non-hemolytic unconjugated hyperbilirubinemia ith glucuronyl transferase deficiency; clinical, biochemical, pharmacologic, and genetic evidence for heterogeneity. Am J Med 47: , Black M, Fevery J, Parker D, et al: Effect of phenobarbitone on plasma [ 14 C ]-bilirubin clearance in patients ith unconjugated hyperbilirubinemia. Clin Sci Molec Med 46:1-17, Wheeler HO, Cranston WI, Meltzer JI: Hepatic uptake and biliary excretion of indocyanine green in the dog. Proc Soc Exper Bioi Med 99:11-14, Whelan G, Hoch J, Combes B: A direct assessment of the importance of conjugation for biliary treatment of sulfobromophthalein sodium. J Lab Clin Med 75: , Habig WH, Pabst MJ, Jakoby WB: Glutathione S-transferases: the first enzymatic step in mercapturic acid formation. J Bioi Chern 249: , Blaschke TF, Berk PD, Scharschmidt BF, et al: Crigler-Najjar syndrome: an unusual course ith development of neurologic damage at age 18. Pediatr Res 8: , Bernstein LH, Ben-Ezzer J, Gartner L, et al: Hepatic intracellular distribution of tritium-labeled unconjugated and conjugated bilirubin in normal and Gunn rats. J Clin Invest 45: , Blaschke TF, Berk PD, Rodkey FL, et al: Drugs and the liver. I. Effects of glutethimide and phenobarbital on hepatic bilirubin clearance, plasma bilirubin turnover and carbon monoxide production in man. Biochem Pharmacol 23: , Black M, Sherlock S: Treatment of Gilbert's syndrome ith phenobarbitone. Lancet 1: , Yoshida A: Hemolytic anemia and G6PD deficiency. Science 179: , Rizzi D, Daniele F, Capurso A, et al: Ittero cronico non emolitico a bilirubina indiretta (sindrome di Gilbert) con accumulo di pigmento nel fegato e ritenzione di BSP. Epatholgia 15:77-88, 1969