Liver cirrhosis : A seven year follow ~ up

TSP TRAN, SV FEINMAN, B BERR1S. Liver cirrhosis: A seven year follow-up. Can J Gastroenterol 1991;5(1):11-14. The status of 121 patients who were found to have liver cirrhosis on liver biopsy in 1981 was assessed seven years later. The etiology of the cirrhosis was alcoholic in 52%, cryptogenic in 29.8%, hepatitis B-related in 9. 1 % and miscellaneous in 9.1 %. In 1981, jaundice was present in 55 patients (45.8%), ascites in 52 (43%), gastrointestinal bleeding in 25 (20.7%) and encephalopathy in 10 (8.3%). During the following seven years an additional 20 patients developed ascites, 15 gastrointestinal bleeding, 32 encephalopathy and three hepatocellular carcinoma. The mortality race was 43.8% at five years and 53.7% at seven years. The principal cause of death was liver failure (40%) , followed by nonliver causes (32.3%) and gastrointestinal bleeding ~ 13.9%). One patient died of hepatocellular carcinoma. Patients who survived seven years had fewer complications when seen in 1981 than those who died during this period (P<0.025). It is concluded that, in Toronto, c irrhosis is often caused by ethanol abuse and hepatitis B infection; that it is associated with significant morbidity and mortality; and that the number of comp! ications when the patient is first seen may be a useful indicator of prognosis. Since many cases of cirrhosis are preventable, the authors suggest chat efforts directed towards prevention of cirrhosis may be more rewarding than chose directed towards therapy.

L IVER CIRRI IOSIS IS A SIGNIFI CANT cause of mortality and morbidity.Despite progress in the medical and surgical therapy of asci tes, porta I hypertension a nd variccal bleeding, five year surviva l rares are srill in the range of 50% when combined etio logies a rc considered ( l-7).The present study was an attempr to characterize a nd assess morbidity ,i nd mortality in a group of patients in whom liver cirrhosis had been diagnosed seven years previously.
The aut hors were able ro trace 12 l of l 76 patients in whom a di agnosis of cirrhosis was made on li ver biopsy in 1981.In th is group the authors studi ed: the d istribution of the varieties of ci rrhosis; the prevalence of complications al the time of initial biopsy anJ over the ensuing seven years; the mortality rate at five and seven years after biopsy; the causes of death; and the relationship of the number of compli cation~ to prognosis.

PATIENTS AND METHODS
Between Febru ary I, l 981 and January 3 1, 1982, 718 patients being investigated for the presence of li ver d isease had diagnostic li ver biopsies in six Toronto teaching hospitals.In 176 patients, biopsies were diagnostic of cirrhosis.Patients with a diagnosis of hepatocellular carcin oma were excluded.These patients were drawn from a larger study of markers and risk facto rs for hepatitis B (8). Information collected included age, gender, etio logy of cirrhosis, complicatio ns, presence of les hemorragies digestives (13,9 % ).Un patient est decede d 'un carc inome hepatocellula ire.Les patients qui ont survecu sept ans avaient moins de complications quanc..l ils one consulte en 198 1 que ceux qui sont decedes durant cette periode (P<0,025 ).On conclut qu'a T oronto la cirrhose est souvent causee par l'e thylisme et l'hepatite virale de type B; qu'elle est associee a une morbidite et a une mortalite significative; et que le nombre de complications relevees lors de la premiere consultat ion pourra it laisser presager le pronostic.Paree que de nombreux cas de c irrhose peuven t etre e vites, les auceu rs suggerent que les efforts invest is cla ns la prevention pourraient s'averer plus fruccueux que ceux qui s'attachent au craitement.History of transfusion 28.9"k markers of hepati tis B and history of transfusion .
Fro m November 1988 co Jan uary 1989 the authors attempted to locate t h e 176 pa ti ents fo r re-study.T he patien t's physician was first contacted.If the physician had lost t rack of t he patient, an attempt was made to contact the patient or the patient's famil y.T he authors were ab le to locate 121 of the 176 patients and obtained info rmation regarding the presence of add itiona l co m p lications such as asc ic es, gastrointestinal bleeding, hepatic encepha lopathy and hepatocell ular carc inoma between 198 l and 1988, as well as the date and cause of an y deaths.ln order to determine whether t he traced group was a valid sample representative of th e entire group, a statistical comparison of the charac te ristics of the traced and lost groups was carried out.Age and gende r distribution were In iti a l comp li catio n s, hepatitis B markers and h istory of transfusion in the traced and lost groups of patients were compared using the Z test.

RESULTS
O n e hun dred and twenty-one patients were t raced, and 55 were lost seven years after t he diagnosis of cirrhosis was made on liver biopsy.Eightye igh t patie n ts were traced through the ir physic ia ns; 18 were contacted directly; a n d 15 we re t raced t h rough t h e ir fa milies.When the c haracteristics of the traced and lost groups identified in 198 1 we re compared, th ere was no statistically significant difference between them (T able I ).T h is suggests tha t t h e traced group was representative of the entire group.
T h e et io logy of c irrhosis in the t raceJ patients was alcoholic in 63 (52%), cryptogcnic in 36 (29.8%), po5r hepati tis B in l l (9.1 %) and miscellaneous in 11 (9.l %).T he distribution of etiologies in the 1Uiscellaneous group was one congestive, three post jejunoileal bypass, one lupoi<l , five primary biliary cirrhosis and on e alph a-I anti• tryps in deficiency.In 1981 the complications were jaund ice in 55 patients, asc ites in 52, and encepha lopathy in 10.
T here was a h istory of gastrointe~tinal bleecl ing in 25 pat ients.T hirty-five patients had no complications, and none h ad evidence of hepatocellular ca rc inoma.Forty patients ha<l no complications, 41 had one, 30 had two, seven had th ree, and two had four.Over t he next seven years an addit ional 20 patie n ts deve loped ascites, IS gast rointesti nal bleeding, and 32 encephalopathy.T hree hepatocellular carcinomas were d iagn osed (Table 2).After seven years 65 patients had died (53.8%).T he five year mortality was 43.8% .lf one assumed chat all 55 patients who were lost to fo llow-up d ied, the seven year mo rta lity rare would be 68.2%.T he seven year mortality was 60.3% in a lcoholic cirrhosis a nd 50% in c ryp toge nic cirrhosis.Three of the 11 pat ients with hepatitis B cirrh osis and six of the 11 patients in t h e mi sce lla n eous gro up <lied.The major causes of death were liver failure ( 40%), nonliver ca uses (cardiovascular, cerebrovascular and nonhepat ic malignanc ies) (32.3%), an<l gascroincescinal bleeding (13.9%).O ne patien t died o( hepatocellular ca rc inoma.In 12.3% of patie nts, the ca use of <leach was unknown . Statistical comparison of the num-   When the present patients were first studied in 198 1, 71. 1% had complica-3 tions of cirrhosis.Jaundice anJ ascites we r e most common, fo llowed by gastrointestinal b leed ing a nJ encephalopath y.The frequency of complicat io n s was ve ry si mila r in two studies reported from Britain ( 1,5 ).
Over the n ext seven years there were numero us add it io n a l complicatio n s, noteJ in T able 2.
Established c irrhosis has a high mortality ra te.In the present patients it was 43.8% at five years and 53.7% at seven years.If it is assumed that all of the patients lost to fo llow-up had died, the seven year mortality rate would be in.creased to 68.2%.In o ther studies the fi ve year morta lity ranges between 30 and 50% (9-11 ).T he mortality rate is very similar in alcoho lic (60.3% ) anJ cryptogenic c irrhosis (50%).The numbers in hepatitis B and in the miscellaneous group of c irrhotics were too sma ll to draw valid conclusions.In o ther studies the mo rtality race appeared to be unrelated to etiology and was very similar in alcoholic, cryptogenic and hepatitis B-re la ted c irrhosis (7)(8)(9).
The major ca uses of d ea th a ttributable to liver disease in the present CAN J GASTROENTEROL VOL 5 No I JANUARY/FEBRUARY 1991 Liver cirrhosis: Seven year follow-up patients were liver failure in 40%, gastrointestina l bl eeding in 13.9% and h epatocellular ca rci nom a in o n e patient.In [2.3% the cause was unknown , and 32.3% died from nonhepatic causes.S imilar findings have been reported from Scandinav ia ( 10), but in Japan there was a much higher mo rtality Jue w hepatocellu lar ca rc inoma (7).In the present study only o ne patient Jied of hepatocellular carc inom a over seven years.The mortali ty rate from hepatocellular carcino ma fo llowing cirrhosis varies greatly.It was o nly 4 1 Yo in the U nited States ( l l) and as high as 44% in Japan; however, this was ove r a 26 yea r fo ll ow-up (7).
H epatocell ular carcinoma is much more prevalem in countries such as Japan, where there is a high prevalence of hepatiti B. It is clea r that the incidence of heparocell ular carcinoma is inc reasing, bo th in the Fa r East (7) and in the West ( l2-l4), possibly because of prolonged surviva l in c irrhotic patients, as well as improved means of detection and awareness of causal relation hips.
T here have been a number of stuJies which have attempted to identify prognostic indicators in c irrhosis.This is now of some importance, s ince it may h e lp to identify patients who will bene fit from a var iety of treatment modalit ies, particularly li ver tra nsplantation.C linical, biochemical and mo rphological facto rs have been assessed often in combination (1 5), and until recently C hikl's criteria have been used most frequently (1 6).The present authors have shown that when the freq uency of initia l complicatio ns in patients who survived was compared to the freq uency in those who had died seven years late r, the re were significantly fewer complicat ions in the survivors.Thus, assessment of the number of complications when patients are first seen may be a usefu l bedside indicator of prognosis, the predictive value of wh ich should be va lidated prospectively in an in.dependentseri es of patients.
In conclusion , the a u thors have sh own that in Toronto, c irrhosis is often ethanol related, cryptogenic or due co chronic hepat itis B infection; that complication s are freque nt; and that mortality is still high .They have also shown that the number of complications when the patient is fim seen may be a useful clinical inJicator of prognosis arn.J may help to iJencify can-diJates for liver transplantation.Since a significant number of cases of c irrhtrb a rc preventable, th e authors suggest that effo rts di rccted cowards prevent ion of ci rrhosis may be more reward ing than those directed towards therapy.
ber of complications per patien t in 1981 between patien ts who survived and those wh o d ied showed that pat ients who~urv i ved had fewer comp lications than those who d ied (P<0.025 ) (FigureI).