Improved prognosis of postoperative hepatocellular carcinoma patients when treated with functional foods: a prospective cohort study

Journl of Heptology 37 (2002) 78 86 www.elsevier.com/locte/jhep Improved prognosis of postopertive heptocellulr crcinom ptients when treted with functionl foods: prospective cohort study Yoichi Mtsui*, Juny Uhr, Sohei Stoi, Mski Kibori, Hitoshi Ymd, Hiroki Kitde, Atsusi Immur, Soichiro Tki, Yusi Kwguchi, A-Hon Kwon, Ysuo Kmiym First Deprtment of Surgery, Knsi Medicl University, 10-15 Fumizono, Moriguchi, Osk 570-8507, Jpn See Editoril, pges 147 150 Bckground/Aims: Active hexose correlted compound (AHCC) is newly developed functionl food. In vitro experiments hve shown tht AHCC enhnces nturl killer cell ctivity, nd my be considered potent biologicl response modifier in the tretment of cncer ptients. However, the effects of AHCC in clinicl setting hve not been reported. We seek to determine whether AHCC cn improve the prognosis of heptocellulr crcinom (HCC) ptients following surgicl tretment. Methods: A prospective cohort study ws performed from Februry 1, 1992 to December 31, 2001. A totl of 269 consecutive ptients with histologiclly confirmed HCC were studied. All of the ptients underwent resection of liver tumor. Time to tretment filure (disese recurrence or deth) nd ten prmeters relted to liver function fter surgery were exmined. Results: Of the 269 ptients, 113 received AHCC orlly fter undergoing curtive surgery (AHCC group). The AHCC group hd significntly longer no recurrence period (hzrd rtio (HR), 0.639; 95% confidence intervl (CI), 0.429 0.952; P ¼ 0.0277) nd n incresed overll survivl rte (HR, 0.421; 95% CI, 0.253 0.701; P ¼ 0.0009) when compred to the control group by Cox s multivrite nlysis. Conclusions: This study suggests tht AHCC intke cn improve the prognosis of postopertive HCC ptients. q 2002 Europen Assocition for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. Keywords: Active hexose correlted compound; AHCC; Biologicl response modifier; Cirrhosis; Functionl food; Heptitis; Heptocellulr crcinom 1. Introduction The incidence of heptocellulr crcinom (HCC) is distributed widely over different geogrphicl res. There is high prevlence of HCC in Asi tht is similr to tht of stomch cncer in Jpn. Moreover, the number of HCC ptients is showing grdul, but definite increse [1]. The prevention nd tretment of the recurrence of HCC following heptic resection hs been studied extensively. Received 8 August 2001; received in revised form 5 Mrch 2002; ccepted 18 Mrch 2002 * Corresponding uthor. Tel.: 181-6-6992-1001, ext. 3262; fx: 181-6- 6992-7343. E-mil ddress: mtsui@tkii.kmu.c.jp (Y. Mtsui). These tretments include repeted heptic resection [2,3], interventionl rdiology (chemoemboliztion) [4,5], percutneous ethnol injections [6,7], percutneous microwve cogultion [8,9], nd the dministrtion of hormonl gents [10 12]. However, the prognosis for HCC remins unstisfctory, with the 5-yer survivl rte fter primry surgicl tretment t pproximtely 40% in Jpn [1]. In ddition to the tretments mentioned bove, there hve been mny ttempts to tret the cncer by stimulting the ptient s immune system. Although severl biologicl response modifiers (BRMs) hve been developed such s BCG, Picibnil, PSK, lentinn, interferon, nd interleukin- 12 [13 16], the clinicl efficcy of these substnces hs not been clerly confirmed. Recently, the efficcy of immu- 0168-8278/02/$20.00 q 2002 Europen Assocition for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. PII: S0168-8278(02)00091-0

Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 79 notherpy in suppressing the postsurgicl recurrence of HCC ws reported s clinicl tril [17]. Active hexose correlted compound (AHCC) is functionl food [18,19] developed by the Amino Up Chemicl Co. Ltd (Spporo, Jpn) in 1989. A food is considered functionl if it is stisfctorily demonstrted to ffect beneficilly one or more trget functions in the body, in wy tht is beyond dequte nutritionl effects nd is relevnt to either the stte of well-being nd helth or the reduction of the risk of disese [20]. The AHCC is n extrct of Bsidiomycotin, which is obtined through the hybridiztion of severl types of mushrooms [21]. Ghoneum et l. reported tht AHCC enhnces the nturl killer (NK) cell ctivity of cncer ptients, nd my be considered potent BRM in the tretment of cncer ptients [22]. It hs been suggested tht NK cell ctivity my be ssocited with cncer incidence [23]. Furthermore, AHCC hs been reported to reduce the metstsis rte of rt mmmry denocrcinoms [21], to increse detoxifiction enzymes in the liver, to protect the liver from CCl 4 -induced liver injury [24], nd to prevent dibetes induced by streptozotocin [25] in niml models. However, there hve been no reports on the effects of AHCC in clinicl setting. This study ws initited to evlute the effects of AHCC, s n orlly dministered BRM, on the prognosis of ptients with HCC following surgicl tretment. 2. Ptients nd methods To determine whether AHCC cn improve the prognosis of HCC ptients following surgicl tretment, prospective cohort study ws performed. All consecutive ptients with HCC who underwent surgicl tretment from Februry 1, 1992 to October 31, 2001 t the First Deprtment of Surgery, Knsi Medicl University, Osk, Jpn were included in this study, if they met the following criteri: (1) the ptient hd undergone curtive resection of their liver tumor t our Deprtment nd (2) the presence of histologiclly proven HCC in their resected liver specimen ws demonstrted. The therpeutic options were offered to ll of the ptients during their hospitliztion. The enrolled ptients were ddressed to ech rm of the study bsed on their choice of the therpeutic options, nd were trusted with the self-dministrtion of AHCC. If the ptient selected the AHCC ingestion, they begn ingesting AHCC t 3.0 g/dy from the dte of their dischrge. The primry endpoint ws survivl nd the secondry endpoint ws no recurrence period. In ddition, ten biochemicl prmeters were exmined yerly to evlute the liver function until deth or the end of the observtion period (December 31, 2001). These prmeters include the serum levels of sprtte trnsminse ctivity (AST), lnine trnsminse ctivity, lkline phosphtse ctivity, g-glutmyltrnsferse ctivity (GGT), totl bilirubin, lbumin, cholinesterse ctivity, pltelet count, - fetoprotein, nd protein induced by vitmin K bsence (PIVKA II). Rndomiztion ws not performed in this study, nd plcebo ws not used for the controls. The study protocol conformed to the ethicl guidelines of our institute nd ws pproved by the institutionl review committee. AHCC ws generously provided by the Amino Up Chemicl Co. Ltd, nd ws developed by extrction from cultured broth of Bsidiomycotin. 2.1. Ptients By October 31, 2001, totl of 269 ptients underwent surgicl tretment for HCC. Of these 269 ptients, totl of 47 cses were excluded s follows: 28 cses of non-curtive resection, four cses of opertive deth, seven cses of hospitl deth, one cse with mentl disorder, one cse with primry biliry cirrhosis, one cse with histologiclly proven combined type of HCC nd cholngiocellulr crcinom, nd five cses tht withdrew from follow-up just fter dischrge. As result, the remining 222 ptients were enrolled in this study nd were observed either until deth or until the lst follow-up dte (December 31, 2001) for the living ptients. Of these 222 ptients, 113 were given AHCC (3.0 g/dy) orlly fter undergoing surgery, in ccordnce with the preferences of the ptient (AHCC group). The dministrtion of AHCC continued until deth or to the lst follow-up dte for the living ptients. The remining 109 ptients were monitored fter the heptectomy, but were not given AHCC (control group). The no recurrence rte nd the overll survivl of the ptients in the AHCC group were compred to tht of the control group. The im of this study ws explined to ll of the pproved ptients in dvnce, nd informed consent ws obtined. All of the ptients were trusted with their choice of AHCC ingestion, following the informed consent. Therefore, the ptients were enrolled in either the AHCC group or the control group entirely ccording to their preferences. Although not controlled study, we obtined similr number of ptients in ech group with the sme clinicl nd pthologicl chrcteristics ccording to the preferences of the ptients. In the erly yers of the study, few more ptients preferred the control group thn the AHCC group. However, the number of ptients who preferred the functionl food grdully incresed. Eventully, pproximtely hlf of the ptients preferred the functionl food t the end of the study period of 9 yers nd 11 months. This chnge in preference resulted in the difference of the medin follow-up period between the two groups shown in Section 3. 2.2. Follow-up Periopertive clinicl prmeters such s the ptient chrcteristics, preopertive liver function dt, opertive fctors, nd tumor chrcteristics were compred between the AHCC nd control groups. Cirrhotic sttus ws histopthologiclly determined in non-cncerous liver tissues ccording to the New Inuym Clssifiction [26]. The stging system used followed the Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer by the Liver Cncer Study Group of Jpn [27], which is commonly used in Jpn. The overll survivl, defined s the intervl between the dte of surgery nd the dte of deth or the lst follow-up informtion for the living ptients, ws lso evluted. The most common cuse of deth ws cncer, but liver filure nd vricel bleeding were included mong the cuses of deth. The no recurrence rte ws lso evluted, nd ws defined s the intervl between the dte of surgery to the dte tht dignosis of recurrence ws confirmed by positive sonogrm, computed tomogrphy, mgnetic resonnce imging, or heptic ngiogrphy. The no recurrence rte ws clculted fter censoring the ptients who hd not shown recurrence t the time of deth. All ptients were given follow-up exmintions with routine liver biochemicl tests. Every 3 months, biochemicl tests were performed t the centrl hospitl lbortory. A liver ultrsound ws lso performed every 3 months. In ddition, computed tomogrphy nd/or mgnetic resonnce imging were performed every 6 months. Finlly, n ngiogrphic exmintion ws performed fter dmission when recurrence ws suspected. Once n intrheptic recurrence hd been confirmed, ptients in both groups generlly received trnsrteril chemo-emboliztion (TACE), wheres some ptients with recurrence underwent lterntive tretments (Tble 1). Ptients without recurrence were not treted with ny other drugs for cncer during follow-up. 2.3. Sttisticl nlysis In order to evlute the homogeneity of the tretment vs. control groups with respect to periopertive clinicl fctors, dt ws nlyzed using the chi-squre test or Mnn Whitney U-test to compre differences between two series. A two-wy nlysis of vrince with Scheffé s F-test ws used

80 Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 Tble 1 Clinicl bckground of ptients treted with AHCC nd controls AHCC group Control group P Ptient chrcteristics Age (yers) 65 (59 70) 63 (57 68) 0.134 Gender (mle/femle) 81/25 87/15 0.148 Cirrhosis (yes/no) 49/57 38/64 0.242 Child clssifiction (A/B) 73/33 77/25 0.363 Alcohol intke (yes/no) 46/60 47/55 0.803 Esophgel vrices (yes/no) 28/52 22/53 0.119 Heptitis virl infection Type B/type C/none 32/85/12 26/84/10 0.785 Preopertive TACE (yes/no) 46/59 45/55 0.975 Previous or concurrent mlignncy (yes/no) 9/97 10/92 0.930 Tretment for recurrence 35/4 57/13 0.382 (yes/no) TACE 35 50 PEIT 1 3 PMCT 0 6 Re-resection 1 6 Systemic chemotherpy 1 1 Irrdition 3 1 Cuse of deth Recurrence/non-recurrence 21/2 47/4 0.999 Preopertive liver function dt Albumin (g/l) 37 (34 40) 39 (36 41) 0.012 b Totl bilirubin (mg/dl) 0.8 (0.6 1.1) 0.8 (0.6 1.1) 0.578 Cholinesterse (U/l) 3721 (3034 4371) 3787 (2824 4851) 0.808 Asprtte trnsminse (U/l) 45 (31 63) 44 (32 61) 0.835 Alnine trnsminse (U/l) 47 (28 72) 43 (26 62) 0.386 Alkline phosphtse (U/l) 243 (183 314) 218 (165 300) 0.966 g-glutmyltrnsferse (U/l) 61 (33 95) 53 (34 87) 0.875 Pltelet count ( 10 9 /l) 115 (91 166) 138 (100 190) 0.012 b Prothrombin time (%) 90 (82 96) 90 (82 100) 0.505 Hepplstin test (%) 83 (73 96) 87 (74 98) 0.349 Antithrombin III (%) 79 (70 87) 81 (69 92) 0.625 ICG R 15 min (%) 16.8 (12.2 23.3) 15.8 (10.6 21.0) 0.100 ICG K vlue 0.122 (0.096 0.147) 0.121 (0.103 0.150) 0.398 Redox tolernce index 0.478 (0.250 0.880) 0.425 (0.185 0.836) 0.576 99m Tc-GSA liver scintigrphy R mx 0.393 (0.330 0.507) 0.422 (0.284 0.500) 0.717 Opertive dt Resected liver volume 0.719 Sub-segmentectomy or less 65 57 More thn subsegmentectomy, 28 30 less thn lobectomy Lobectomy or more 13 15 Totl blood loss (ml) 1006 (475 1593) 920 (420 1710) 0.635 Blood trnsfusion (yes/no) 51/55 52/50 0.784 Opertion length (min) 270 (205 340) 263 (200 345) 0.600 Postopertive complictions 28/78 23/79 0.627 (yes/no) Postopertive hospitl sty (dys) 22 (17 34) 24 (18 32) 0.431 Tumor chrcteristics Tumor dimeter (cm) 3.0 (2.0 4.5) 3.1 (2.2 5.0) 0.411 Number of nodules (single/two 81/25 76/26 0.874 or more) Differentition (well/moderte/ 12/72/19 9/66/22 0.694 poor) Cpsule (yes/no) 89/17 83/16 0.981

Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 81 Tble 1 (continued) AHCC group Control group P Portl vein thrombosis (yes/no) 5/101 3/99 0.760 Stge (I/II/III/IVA) c 24/49/26/7 20/50/21/11 0.633 Bsl -fetoprotein (Less thn 100 mg/l/more) 65/41 61/40 0.999 Bsl PIVKA II (Less thn 100 AU/l/more) 44/54 43/55 0.999 AHCC, ctive hexose correlted compound; TACE, trnsrteril chemoemboliztion; PEIT, percutneous ethnol injection therpy; PMCT, percutneous microwve cogultion therpy; ICG, indocynine green; GSA, glctosyl humn serum lbumin; PIVKA, protein induced by vitmin K bsence. b Significnt. c The Stge ws defined ccording to the Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer by the Liver Cncer Study Group of Jpn. The dt re expressed s medin (interqurtile intervl). to compre the postopertive course of the lbortory dt between the two groups. The no recurrence curves nd the overll survivl curves were plotted by the Kpln Meier method, nd log rnk tests were lso performed. The time-fixed Cox s proportionl hzrd model ws used to estimte the effects of AHCC on the no recurrence rte nd the overll survivl. For univrite screening purposes, of the 40 potentil risk fctors shown in Tble 1, the tretment for recurrence nd the cuse of deth were excluded nd the remining 38 fctors were exmined univritely by the Cox s model, becuse these two vribles were time-dependent fctors. These 38 vribles were ll ctegorized s binry. Cirrhotic sttus ws dichotomized into two ctegoricl dt group: with histopthologiclly confirmed cirrhosis or without histopthologiclly confirmed cirrhosis. The stging system [27] ws divided into two ctegoricl groups: I/II nd III/IVA. The resected liver volume ws divided into two ctegoricl groups: sub-segmentectomy or less nd more thn sub-segmentectomy. Since the medin of the mount of lcohol intke ws 0 g/dy of ethnol, the mount of lcohol intke ws dichotomized into two groups: ptients with drinking hbit nd those without. The number of nodules ws seprted into two ctegoricl dt: single or not. The levels of tumor mrkers (-fetoprotein nd PIVKA II) tht were highly skewed were divided into two ctegoricl groups t 100 mg/l nd 100 AU/l, respectively, which were cliniclly relevnt. Other continuous vribles were dichotomized into two groups t their overll medin. All fctors found to be significntly ssocited univritely with survivl were included in the multivrite Cox s nlysis with stepwise method. The ssumption of the proportionl hzrds ws checked using the log log plotting method nd the prllel lines between the two groups were confirmed. A P-vlue of less thn 0.05 ws considered to be sttisticlly significnt. 3. Results The use of AHCC showed no side effects. Only three ptients in the AHCC group refused to continue the use of AHCC during the study due to slight nuse. These three cses were censored t tht time. Some cses hd minor complints of difficulty in swllowing the AHCC due to the grnulr type of its mteril. However, these ptients did not stop the tretment. Four ptients in the control group begn to tke AHCC during the observtion period becuse they chose to tke AHCC. These four cses were censored t tht time. Tble 1 demonstrtes the similr clinicl bckgrounds of the ptients between the two tretment groups. The lbumin levels nd the pltelet count were significntly different between the two groups preopertively. However, the differences were disdvntgeous to the AHCC group. Most ptients were dignosed with n underlying virl heptitis or cirrhosis, but they lso hd well-compensted liver function. No ptients hd scites preopertively. 3.1. No recurrence rte nd overll survivl By December 31, 2001, 39 (34.5%) ptients hd recurrences of HCC in the AHCC group, while 72 (66.1%) hd recurrences in the control group. The results suggest tht the use of AHCC hd significnt effect (P ¼ 0:0335, log rnk test) on the no recurrence rte (Fig. 1A). Only 23 (20.4%) ptients hd died in the AHCC group by the end of the followup period, wheres 51 (46.8%) hd died in the control group t the end of the follow-up period. The cuses of deth were 91.3% recurrence in the AHCC group nd 92.2% in the control group. Ptient survivl ws significntly higher (P ¼ 0:0032, log rnk test) in the AHCC group (Fig. 1B). The follow-up period rnged from 2 to 108 months in the AHCC group, nd from 2 to 117 months in the control group. The medin follow-up period ws 28 months in the AHCC group nd 30 months in the control group. Time-fixed Cox s univrite nlysis ws performed using ll of the 38 vribles mentioned bove (Tble 1). Of these 38 vribles, the following 11 vribles were significntly relted univritely to the no recurrence rte: AHCC intke, cirrhosis, bsl totl bilirubin, bsl cholinesterse ctivity, bsl ntithrombin III ctivity, ICG R 15 min (%), blood trnsfusion, number of nodules, Stge, bsl - fetoprotein levels, nd bsl PIVKA II levels (Tble 2). These vribles were included in the Cox s multivrite nlysis. In the lst step, the following five vribles entered the model nd could not be removed: AHCC intke, bsl totl bilirubin, bsl cholinesterse ctivity, number of nodules, nd bsl -fetoprotein levels (Tble 4). Accordingly, these five vribles were significntly ssocited with the no recurrence rte, nd were found to be independent fctors. In the multivrite nlysis, the hzrd rtio of no recurrence in the AHCC group ws reduced to 0.639 from 0.658 by the univrite nlysis (Tble 4).

82 Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 Of the 38 vribles, the following 13 were significntly relted univritely to the overll survivl: AHCC intke, cirrhosis, Child clssifiction, bsl lbumin level, bsl cholinesterse ctivity, ICG 15 min (%), ICG K vlue, opertion length, number of nodules, blood trnsfusion, Stge, bsl -fetoprotein levels, nd bsl PIVKA II levels (Tble 3). These 13 vribles were included in the Cox s multivrite nlysis. In the lst step, the following five vribles entered the model nd could not be removed: AHCC intke, ICG 15 min (%), number of nodules, Stge, nd bsl -fetoprotein levels (Tble 4). Accordingly, these five vribles were significntly ssocited with the overll survivl, nd were found to be independent fctors. In the multivrite nlysis, the hzrd rtio of overll survivl in the AHCC group ws reduced to 0.421 from 0.485 by the univrite nlysis (Tble 4). 3.2. Biochemicl prmeters Ten biochemicl prmeters were investigted for period of 5 yers fter surgery in the two groups. Of these ten prmeters, three prmeters, including the serum levels of AST, GGT, nd cholinesterse ctivity, were significntly improved in the AHCC group thn in the controls, s demonstrted using two-wy nlysis of vrince (Fig. 2). No significnt differences were observed in the other seven prmeters, which included the serum levels of lnine trnsminse ctivity, lkline phosphtse ctivity, totl bilirubin, lbumin levels, pltelet count, -fetoprotein levels, nd PIVKA II levels (dt not shown). To eliminte the potentil for tumor relted effect on these ten biochemicl prmeters, nd to clrify whether the AHCC improved liver function independently from the tumor, the prmeters were lso investigted fter excluding the dt of the ptients who hd recurrence. Consequently, of the ten prmeters mentioned bove, the sme three prmeters, including the serum levels of AST, GGT, nd cholinesterse ctivity, were significntly improved in the AHCC Tble 2 Significnt vribles in the univrite nlysis for the no recurrence rte Significnt vribles b SE HR(95% CI) P Fig. 1. Kpln Meier estimtes of the no recurrence rte nd overll survivl of HCC ptients fter heptic resection. The thick line indictes survivl in the AHCC group, nd the thin line represents the control group. (A) No recurrence rte. There ws significnt difference between the two groups on the log rnk test (P ¼ 0.0335). (B) Overll survivl. There ws lso significnt difference between the two groups (P ¼ 0.0032). AHCC intke 20.418 0.200 0.658 0.0369 (Yes vs. no) (0.444 0.975) Cirrhosis 20.485 0.194 0.616 0.0126 (No vs. yes) (0.421 0.901) Bsl totl bilirubin 20.461 0.192 0.631 0.0164 (,0.8 vs. $0.8 mg/dl) (0.433 0.919) Bsl cholinesterse 20.502 0.196 0.605 0.0102 ($3768 vs.,3768 U/l) (0.412 0.888) Bsl ntithrombin III 20.523 0.222 0.593 0.0187 ($80 vs.,80%) (0.383 0.917) ICG R 15 min 20.396 0.193 0.673 0.0398 (,16 vs. $16%) (0.461 0.982) Blood trnsfusion 20.432 0.193 0.649 0.0255 (No vs. yes) (0.445 0.948) Number of nodules 20.980 0.199 0.375,0.0001 (Single vs. two or more) (0.254 0.554) Stge 20.550 0.195 0.577 0.0048 (I/II vs. III/IVA) b (0.394 0.846) Bsl -fetoprotein 20.767 0.209 0.464 0.0002 (,100 vs. $100 mg/l) (0.308 0.699) Bsl PIVKA II 20.626 0.203 0.535 0.0021 (,100 vs. $100 AU/l) (0.359 0.797) AHCC, ctive hexose correlted compound; ICG, indocynine green; PIVKA, protein induced by vitmin K bsence; b, regression coefficients; SE, stndrd error; HR, hzrd rtio; CI, confidence intervl. b The Stge ws defined ccording to the Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer by the Liver Cncer Study Group of Jpn.

Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 83 group thn in the controls (Fig. 3), wheres no significnt differences were observed in the other seven prmeters. 4. Discussion HCC is mjor helth concern worldwide, with n incidence of pproximtely one million cses per yer [28]. Recently, the erly detection of HCC hs become possible becuse of progress in dignostic imging, nd the incidence of resection for HCC hs incresed gretly during the lst decde. As result, the short-term outcome hs improved gretly, nd no-mortlity series on liver resection for HCC were reported [29]. Furthermore, there hve been significnt improvements in ptients prognosis for those cses with HCC who were treted recently with liver resection in comprison to those treted with resection in the erly 1990s [30]. However, the long-term results re not yet stisfctory. Although heptic resection is the most effective form of tretment for ptients with HCC, the incidence of postopertive recurrence, which is the min cuse of the poor long-term results, remins extremely high [31]. Moreover, the cumultive intrheptic recurrence rte hs Tble 3 Significnt vribles in the univrite nlysis for the overll survivl Significnt vribles b SE HR (95% CI) P AHCC intke 20.723 0.252 0.485 0.0041 (Yes vs. no) (0.296 0.795) Cirrhosis 20.509 0.232 0.601 0.0278 (No vs. yes) (0.382 0.946) Child clssifiction 20.596 0.254 0.551 0.0188 (A vs. B) (0.335 0.906) Bsl lbumin 20.503 0.241 0.605 0.0370 ($38 vs.,38 g/l) (0.377 0.970) Bsl cholinesterse 20.534 0.242 0.586 0.0276 ($3768 vs.,3768 U/l) (0.365 0.943) ICG R 15 min 20.475 0.240 0.622 0.0480 (,16 vs. $16%) (0.388 0.996) ICG K vlue 20.545 0.258 0.580 0.0347 ($0.122 vs.,0.122) (0.350 0.962) Opertion length 20.506 0.237 0.603 0.0326 (,265 vs. $265 min) (0.379 0.959) Number of nodules 20.946 0.233 0.388,0.0001 (Single vs. two or more) (0.246 0.613) Blood trnsfusion 20.578 0.244 0.561 0.0177 (No vs. yes) (0.348 0.905) Stge 20.670 0.235 0.512 0.0044 (I/II vs. III/IVA) b (0.323 0.812) Bsl -fetoprotein 21.055 0.242 0.348,0.0001 (,100 vs. $100 mg/l) (0.217 0.560) Bsl PIVKA II 20.586 0.238 0.556 0.0138 (,100 vs. $100 AU/l) (0.349 0.887) AHCC, ctive hexose correlted compound; ICG, indocynine green; PIVKA, protein induced by vitmin K bsence; b, regression coefficients; SE, stndrd error; HR, hzrd rtio; CI, confidence intervl. b The Stge ws defined ccording to the Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer by the Liver Cncer Study Group of Jpn. been reported t 100% t 5 yers fter the resection of single HCC in cirrhotic ptients with virl heptitis [32]. To prevent recurrence nd/or to prolong survivl, the most widely used option is djuvnt chemotherpy through ctheter inserted into the heptic rtery [4]. However, the efficcy of these gents is very poor, the incidence of side effects is high, nd there is no cler evidence suggesting tht their dministrtion results in improved survivl [28]. Furthermore, therpeutic doses of nti-cncer drugs hve been reported to reduce the host nti-tumor immune response, nd the postopertive use of immunosuppressnts hs been shown to ccelerte the recurrence of mlignncy [33]. Thus, the serch for other potentilly useful therpeutic pproches is necessry. Recently, Tkym et l. reported some efficcy of doptive immunotherpy on HCC recurrence on the bsis of rndomized clinicl tril [17]. The remrkble results shown in this tril were very interesting nd encourging, lthough the procedure is reltively complicted nd time-consuming, nd requires hospitliztion to perform. Under these circumstnces, other options such s immunotherpy or rdition hve little prcticl ppliction in dily clinicl setting, nd hve been used only within reserch trils [28]. The disppointing stte of medicl tretment for HCC justifies the interest in the dministrtion of functionl foods such s AHCC s Tble 4 Multivrite nlysis for the no recurrence rte nd the overll survivl Significnt vribles b SE HR (95% CI) P No recurrence AHCC intke 20.448 0.203 0.639 0.0277 (Yes vs. no) (0.429 0.952) Bsl totl bilirubin 20.443 0.196 0.642 0.0242 (,0.8 vs. $0.8 mg/dl) (0.437 0.944) Bsl cholinesterse 20.440 0.199 0.644 0.0267 ($3768 vs.,3768 U/l) (0.436 0.950) Number of nodules 20.880 0.203 0.415,0.0001 (Single vs. two or more) (0.279 0.618) Bsl -fetoprotein 20.561 0.215 0.571 0.0092 (,100 vs. $100 mg/l) (0.374 0.870) Overll survivl AHCC intke 20.865 0.260 0.421 0.0009 (Yes vs. no) (0.253 0.701) ICG 15 min 20.728 0.249 0.483 0.0035 (,16 vs. $16%) (0.296 0.787) Number of nodules 20.807 0.258 0.446 0.0017 (Single vs. two or more) (0.269 0.739) Stge 20.583 0.266 0.558 0.0285 (I/II vs. III/IVA) b (0.331 0.941) Bsl -fetoprotein 21.010 0.248 0.364,0.0001 (,100 vs. $100 mg/l) (0.224 0.592) AHCC, ctive hexose correlted compound; ICG, indocynine green; PIVKA, protein induced by vitmin K bsence; b, regression coefficients; SE, stndrd error; HR, hzrd rtio; CI, confidence intervl. b The Stge ws defined ccording to the Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer by the Liver Cncer Study Group of Jpn.

84 Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 Fig. 2. Biochemicl prmeters in HCC ptients fter heptic resection. The closed circles indicte the levels of the pproprite prmeters in the AHCC group, nd the open circles indicte the levels in the controls. There were significnt differences between the two groups on two-wy nlysis of vrince with Scheffé s F-test. The dt re expressed s medins. (A) Asprtte trnsminse ctivity, P ¼ 0.0011; (B) g-glutmyltrnsferse ctivity, P, 0.0001; (C) cholinesterse ctivity, P ¼ 0.0013. Fig. 3. Biochemicl prmeters in HCC ptients who hd no recurrence fter heptic resection. The closed circles indicte the levels of the pproprite prmeters in the AHCC group, nd the open circles indicte the levels in the controls. There were significnt differences between the two groups on two-wy nlysis of vrince with Scheffé s F-test. The dt re expressed s medins. (A) Asprtte trnsminse ctivity, P ¼ 0.0176; (B) g-glutmyltrnsferse ctivity, P ¼ 0.0007; (C) cholinesterse ctivity, P ¼ 0.0266.

Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 85 BRM, lthough its nti-tumor effects remin uncertin in clinicl setting. A mjor disdvntge in this study is tht it is not rndomized. A rndomized tril would be of higher vlue, since the type of tretment lloction we hve followed my prompt severe bises tht re very difficult to control. However, most of the functionl foods including AHCC, re on the mrket in Jpn nd re vilble esily without prescription becuse it is not medicine tht is required to be prescribed by physicin. It is very difficult to strictly control the ptients ddressed to ech rm, becuse the ptients who prefer the functionl food my obtin it out of the tril. Therefore, it is difficult to complete rndomized tril in regrd to the functionl foods. Alloction to ech rm on the bsis of the ptient s own selection rther thn rndomiztion my be better for the tril of functionl foods. In the cse of the functionl food, the ptients enrolled my divide into ech rm more exctly in the tril ccording to the preferences of the ptients, compred to those in the rndomized tril. Therefore, rndomiztion ws not considered in this study, despite mny issues for severe bises in non-rndomized trils. AHCC is n extrct obtined from severl species of mushrooms. AHCC contins vrious components, but the ctive component is n oligoscchride with n verge moleculr weight of pproximtely 5000 [21]. Interestingly, in contrst to conventionl ctive components such s the b- 1,3-glucn structurl component found in PSK nd lentinn, the glucose oligomer in AHCC hs n -1,4-linkge structure nd some esterified hydroxy groups [21]. However, AHCC my function s BRM in the sme mnner s PSK nd lentinn. In vitro experiments [21] hve shown tht AHCC restores the NK cell ctivity tht ws depressed by n nti-cncer drug, nd stimulted peritonel mcrophge cytotoxicity, NO production, nd cytokine production. The combintion of the nti-cncer drug nd AHCC significntly improved the prognosis of mice fter the excision of their primry tumors. Both NK cells nd mcrophges hve been reported to be involved in the inhibition of tumor metstsis following ctivtion by BRMs [21]. Therefore, this AHCC effect my be medited by the nturl host immunity, which is restored or ctivted by AHCC. These findings suggest tht AHCC my induce its therpeutic effects on the survivl of HCC ptients s result of NK cell nd mcrophge ctivtion. Accordingly, AHCC should be considered s potent BRM, nd its nti-cncer ctivity my be medited through host immunomodultion. Recently, AHCC ws reported to protect the liver from CCl 4 -induced liver injury in n niml model [24]. The incresed survivl rte of the AHCC group suggests tht AHCC my hve hd beneficil effects on the clinicl course of ptients with heptitis or cirrhosis, in ddition to its nti-cncer effects. Indeed, AHCC intke seemed to improve the heptitis disese stte, s suggested by improvements in the postopertive levels of AST nd GGT reported here. In ddition, the observtion tht the cholinesterse ctivity incresed in the AHCC group suggests tht AHCC intke results in some nutritionl improvements in these ptients. The nlysis ws lso performed in ptients who hd no recurrence. This would eliminte the potentil for tumor relted effect. Consequently, improvements of heptitis nd nutritionl sttus in those who hd no recurrence were lso shown. These results show tht the AHCC my improve liver function independent from the tumor. However, cution must be observed in interpreting the evlution of these results. These biochemicl prmeters my improve becuse only those tht survive cn be exmined. This elimintes those ptients with poorer bseline profiles. In ddition, if the control group hs n undetected bis towrd worse outcome, the difference might be rtificil. Our results suggest tht the use of AHCC decreses both the probbility of recurrence nd the hzrd of deth due to HCC nd/or liver cirrhosis. Furthermore, AHCC intke might lso improve heptitis or cirrhosis in the ptients who hd no HCC. The improvements in liver function in the AHCC group pper to reflect n improved prognosis, lthough rndomized-controlled tril is needed to confirm this observtion if possible. The mechnisms responsible for the nti-cncer ctivity nd/or the heptitis-ttenuting effect of AHCC were not explored in this study. At present, the effects of AHCC s the result of single ingredient re difficult to explin, nd it is similrly difficult to rech ny conclusion regrding the complex effects of AHCC on ptient survivl. AHCC intke resulted in improved liver function, the prevention of recurrence of HCC fter resection, nd the prolonged survivl of postopertive HCC ptients without ny dverse effects. Therefore, AHCC tretment is vluble djuvnt therpy s BRM in these ptients. If possible, these observtions need to be confirmed in lrger, rndomized-controlled double-blind trils. In ddition, more detiled studies re required to elucidte the mechnisms responsible for the effects of AHCC. Acknowledgements We thnk Kohji Wkme nd Kenichi Kosun (Amino Up Chemicl Co. Ltd) for providing the AHCC free of cost. References [1] The Liver Cncer Study Group of Jpn. Primry liver cncer in Jpn. In: Wtnbe S, Toming S, Kkizoe T, editors. Jpnese Cncer Assocition. Cncer Tretment nd Survivl, Gnn Monogrph on Cncer Reserch No. 43, Tokyo: Jpn Scientific Societies Press, 1995. pp. 81 95. [2] Shimd M, Tkenk K, Tguchi K, Fujiwr Y, Gion T, Kjiym K, et l. Prognostic fctors fter repet heptectomy for recurrent heptocellulr crcinom. Ann Surg 1998;227:80 85. [3] Shimd M, Mtsumt T, Tketomi A, Ymmoto K, Itsk H,

86 Y. Mtsui et l. / Journl of Heptology 37 (2002) 78 86 Sugimchi K. Repet heptectomy for recurrent heptocellulr crcinom. Surgery 1994;115:703 706. [4] Roul JL, Guyder D, Bretgne JF, Heutot JF, Duvuferrier R, Bourguet P, et l. Prospective rndomized tril of chemoemboliztion versus intr-rteril injection of 131 I-lbeled-iodized oil in the tretment of heptocellulr crcinom. Heptology 1997;26:1156 1161. [5] Mjno PE, Adm R, Bismuth H, Csting D, Ariche A, Krisst J, et l. Influence of preopertive trnsrteril lipiodol chemoemboliztion on resection nd trnsplnttion for heptocellulr crcinom in ptients with cirrhosis. Ann Surg 1997;226:688 703. [6] Cstellno L, Clndr M, Blnco CV, De Sio I. Predictive fctors of survivl nd intrheptic recurrence of heptocellulr crcinom in cirrhosis fter percutneous ethnol injection: nlysis of 71 ptients. J Heptol 1997;27:862 870. [7] Ebr M, Ohto M, Sugiur N, Kit K, Yoshikw M, Okud K, et l. Percutneous ethnol injection for the tretment of smll heptocellulr crcinom. Study of 95 ptients. J Gstroenterol Heptol 1990;5:616 626. [8] Seki T, Wkbyshi M, Nkgw T, Itho T, Shiro T, Kunied K, et l. Ultrsoniclly guided percutneous microwve cogultion therpy for smll heptocellulr crcinom. Cncer 1994;74:817 825. [9] Murkmi R, Yoshimtu S, Ymshit Y, Mtsukw T, Tkhshi M, Sgr K. Tretment of heptocellulr crcinom: vlue of percutneous microwve cogultion. AJR 1995;164:1159 1164. [10] Kouroumlis E, Skordilis P, Thermos K, Vsilki A, Moschndre J, Mnousos ON. Tretment of heptocellulr crcinom with octreotide: rndomised controlled study. Gut 1998;42:442 447. [11] Cncer of the Liver Itlin Progrmme Group. Tmoxifen in tretment of heptocellulr crcinom: rndomized controlled tril. Lncet 1998;352:17 20. [12] Pignt S, Dniele B, Gllo C, De Vivo R, Monfrdini S, Perrone F. Endocrine tretment of heptocellulr crcinom. Any evidence of benefit?. Eur J Cncer 1998;34:25 32. [13] Mizushim Y, Yuhki N, Hosokw M, Kobyshi H. Diminution of cyclophosphmide-induced suppression of ntitumor immunity by n immunomodultor PS-K nd combined therpeutic effects of PS-K nd cyclophosphmide on trnsplnted tumor in rts. Cncer Res 1982;42:5176 5180. [14] Nishiok Y, Hiro M, Robbins PD, Lotze MT, Thr H. Induction of systemic nd therpeutic ntitumor immunity using intrtumorl injection of dendritic cells geneticlly modified to express interleukin 12. Cncer Res 1999;59:4035 4041. [15] Akiym J, Kwmur T, Gotohd E, Ymd Y, Hosokw M, Kodm T, et l. Immunochemotherpy of trnsplnted KMT-17 tumor in WKA rts by combintion of cyclophosphmide nd immunostimultory protein-bound polyscchride isolted from bsidiomycetes. Cncer Res 1977;37:3042 3045. [16] Nkzto H, Koike A, Sji S, Ogw N, Skmoto J. Efficcy of immunochemotherpy s djuvnt tretment fter curtive resection of gstric cncer. Lncet 1994;343:1122 1126. [17] Tkym T, Sekine T, Mkuuchi M, Ymski S, Kosuge T, Ymmoto J, et l. Adoptive immunotherpy to lower postsurgicl recurrence rtes of heptocellulr crcinom: rndomised tril. Lncet 2000;356:802 807. [18] Milner JA. Functionl foods: the US perspective. Am J Clin Nutr 2000;71(Suppl):1654S 1659S. [19] Roberfroid MB. Concepts nd strtegy of functionl food science: the Europen perspective. Am J Clin Nutr 2000;71(Suppl):1660S 1664S. [20] Agett PJ, Alexnder J, Alles M, Anderson PA, Antonie JM, Ashwell M, et l. Scientific concepts of functionl foods in Europe consensus document. Br J Nutr 1999;81:S1 S27. [21] Mtsushit K, Kurmitsu Y, Ohiro Y, Obr M, Kobyshi M, Li YQ, et l. Combintion therpy of ctive hexose correlted compound plus UFT significntly reduces the metstsis of rt mmmry denocrcinom. Anticncer Drugs 1998;9:343 350. [22] Ghoneum M, Ninomiy Y, Torbi M, Gill G, Wojdni A. Active hemicellulose compound (AHCC) enhnce NK cell ctivity of ged mice in vivo. FASEB J 1992;6:A1213 (Abstrct). [23] Imi K, Mtsuym S, Miyke S, Sug K, Nkchi K. Nturl cytotoxic ctivity of peripherl-blood lymphocytes nd cncer incidence: n 11-yer follow-up study of generl popultion. Lncet 2000;356:1795 1799. [24] Sun B, Wkme K, Mukod T, Toyoshim A, Knzw T, Kosun K. Protective effects of AHCC on crbon tetrchloride induced liver injury in mice. Nt Med 1997;51:310 315. [25] Wkme K. Protective effects of ctive hexose correlted compound (AHCC) on the onset of dibetes induced by storeptozotocin in the rt. Biomed Res 1999;20:145 152. [26] Ichid F, Tsuji T, Omt M, Ichid T, Inoue K, Kmimur T, et l. New Inuym Clssifiction: new criteri for histologicl ssessment of chronic heptitis. Int Heptol Commun 1996;6:112 119. [27] The Liver Cncer Study Group of Jpn. The Generl Rules for the Clinicl nd Pthologicl Study of Primry Liver Cncer. 4th ed. Tokyo: Knehr, 2000. [28] Simonetti RG, Liberti A, Angiolini C, Pgliro L. Review. Tretment of heptocellulr crcinom: systemtic review of rndomized controlled trils. Ann Oncol 1997;8:117 136. [29] Torzilli G, Mkuuchi M, Inoue K, Tkym T, Skmoto Y, Sugwr Y, et l. No-mortlity liver resection for heptocellulr crcinom in cirrhotic nd noncirrhotic ptients. Arch Surg 1999;134:984 992. [30] Poon RT-P, Fn S-T, Lo C-M, Ng IO-L, Liu C-L, Lm C-M, et l. Improving survivl results fter resection of heptocellulr crcinom: prospective study of 377 ptients over 10 yers. Ann Surg 2001;234:63 70. [31] Shirbe K, Knemtsu T, Mtsumt T, Adchi E, Akzw K, Sugimchi K. Fctors linked to erly recurrence of smll heptocellulr crcinom fter heptectomy: univrite nd multivrite nlyses. Heptology 1991;14:802 805. [32] Belghiti J, Pnis Y, Frges O, Benhmou JP, Fekete F. Intrheptic recurrence fter resection of heptocellulr crcinom complicting cirrhosis. Ann Surg 1991;214:114 117. [33] Mihich E. Immnosuppression in cncer therpeutics. Trnsplnt Proc 1975;7:275 278.