Gut Supplement, 1991, S 116-S 122

Cystic disease of the liver and biliary tract

A Forbes, I M Murray-Lyon

AbstractThe widespread availability of ultrasoundimaging has led to more frequent recognitionof cystic disease affecting the liver and biliarytract. There is a wide range ofpossible causes.Cystic disease of infective origin is usuallycaused by an Echinococcal species, or as thesequel of a treated amoebic or pyogenicabscess. The clinical and radiological featuresare often then distinctive and will not be dweltupon in this review, except in respect of theircontribution to the differential diagnosis ofnon-infective disorders. The principal non-infective cysts can be conveniently dividedbetween the simple cyst, the polycystic syn-dromes (usually with coexistent renal disease),Caroli's syndrome, and choledochal cysts.The overlap between constituent members ofthese groups, and the association of cysticdisease with hepatic fibrosis (especially withcongenital hepatic fibrosis) has attracted con-siderable attention, and it has been suggestedthat they may all be considered to belongto a hepatobiliary fibrocystic continuum. Inaddition there are a variety of cystic neoplasmsand a miscellany of unusual forms.

Gastrointestinal Unit,Charing Cross Hospital,LondonA ForbesI M Murray-LyonCorrespondence to:Dr I M Murray-Lyon,Gastrointestinal Unit, CharingCross Hospital, LondonW6 8RF.

Investigation and diagnosisLarge cystic lesions may be recognised clinically,but in most cases will be found at ultrasoundexamination undertaken because of symptoms,or laboratory screening tests suggestive ofhepato-biliary disease; cysts are not infrequently anincidental finding when scans are done for un-related clinical problems.An accurate differential diagnosis will often be

possible from ultrasonography alone.'2 Mostcommonly, one or more clearly defined echo-lucent spaces are identified within the liversubstance. When these have thin walls andreduce the signal from more distant structures,the most likely diagnosis is of simple cyst andfurther investigation is usually unnecessary. It issuggested, however, that magnetic resonance(MR) imaging may be particularly helpful indoubtful cases.34 Thicker walled single or mul-tiple cysts with a multilayer or laminated marginare likely to be hydatid, and the presence ofdaughter cysts within the lesion is virtuallypathognomonic. A supportive history and posi-tive serology will usually allow confirmation ofthis diagnosis. Aspiration ofpresumptive hydatidcysts is contra-indicated because of the perceivedhigh risk of anaphylaxis if cyst contents spill intothe peritoneum, and there is usually no need toconsider aspiration if hydatid disease is suspec-ted. The hazard has perhaps been over empha-sised, however, and if hydatid disease is unlikelyand cysts have inconclusive sonographic appear-

ances, aspiration for microbiological and cyto-logical examination is warranted. Several reports- (for example,5 and our own unpublished obser-vations) of needle diagnosis of unsuspected hy-datid disease, and even therapy by ultrasoundguided transcutaneous injection of sclerosant,6 7indicate that if the transhepatic route is taken therisk ofmorbidity is low.

Distinction of abscess from cyst is relativelysimple if an abscess has viscous echo densecontents with a thick wall and densely com-pressed surrounding hepatic parenchyma. Per-cutaneous aspiration allows confirmation of thediagnosis, provides material for microbiologicalexamination, and may be of major therapeuticbenefit. Positive blood cultures or amoebicserology may, however, render aspiration super-fluous, given that small single abscesses can beeffectively managed with systemic antimicro-bials alone. Open surgical drainage remainsnecessary for optimal results in some patientswith chronic multiloculated abscesses.

Cystic lesions in continuity with the biliarytree or with vascular structures may be identifiedby ultrasound (the latter particularly so withDoppler probes), but it will usually be necessaryto obtain contrast studies (by ERCP or angio-graphy) for complete diagnosis. In this contextcomputed tomography and MR images of thehepatobiliary region do not appear at present tooffer many advantages over high resolution sono-graphic examination. The interested reader isreferred to one of a number of illustrated reviewsin the radiological literature. i2I48

Liver biopsy tends to be avoided in theinvestigation of cystic disorders, a stance cer-tainly justified when there is suspicion ofhydatiddisease, and probably so when imaging suggestscholedochal cyst, but the recognition of under-lying fibrosis is then less likely, and the clinicianis denied potentially important prognosticinformation.

SIMPLE HEPATIC CYSTSBenign non-parasitic cysts of the liver, onceconsidered rare because they are generally smalland asymptomatic, are more commonly shownby modem imaging (prevalence between 0 19and 2.5%),2 more so in women (female:male ratio1 .5),2 more often in the right lobe, and probablyofcongenital origin. The cysts are lined by biliarycolumnar epithelium, but it is relatively unusualfor the fluid contents to be bile. If the typicalappearances at sonography described above areseen, no further investigation or treatment isrequired. The sensitivity of aspiration cytologyin potentially neoplastic cysts with no definitesolid component is low, but may be improved byassay ofcyst contents for tumour markers such as

S116

Cystic disease ofthe liverand biliary tract

carcinoembryonic antigen.'0 Other than fromserial scans begun at the time of infection it is notclear that postinfective cysts - longterm sequelaeof unaspirated abscesses - can accurately bedistinguished from congenital benign cysts.' 8

All benign cysts appear subject to a variety ofcomplications. Bleeding into or infection of thecyst contents may be responsible for the onset ofpain or other symptoms; enlargement of cystsalso occurs without obvious cause, and may thenlead to the presence of a painful swelling,"obstructive jaundice'2 or intraperitoneal perfora-tion.'3 Although surgical intervention maybecome necessary, in the absence of peritonitisor severe intraabdominal bleeding, conservativemanagement is probably to be preferred,'4 15with'2 or without'6 percutaneous aspiration.Recurrent symptoms may be prevented by in-ternal drainage ofthe cyst'7 or by the instillation ofsclerosant,'891 the latter course in our view to befollowed only after very careful direct cholangio-graphy has excluded communication with thebiliary tree, if biliary sclerosis is to be avoided.Although there does not seem to be a majorproblem with alcohol injection in analogousmanagement of hydatid cysts7 longterm followup data are sparse. Surgical deroofing of cystswithout drainage ofthe residual cavity is stronglyrecommended20 if surgery is thought indicated.Concern that simple cysts represent a premalig-

nant state has been generated by a series ofreports of neoplasia in apparently preexistingbenign cysts. The tumours are, however, ofdisparate histological type - for example, muco-epidermoid carcinoma,2' multifocal papillarycystadenocarcinoma,22 squamous cell car-cinoma23 - and these may represent coincidentalassociations rather than complications. Theobservation that foci of cystadenoma and cysta-denocarcinoma (the most common malignancydescribed with simple cysts) coexisted withhamartomatous bile ducts24 tends, however, tosupport the view22 that there is a causative link.Although these tumours appear to have a gener-ally poor prognosis, their rarity does not seem tous to warrant indefinite surveillance of asympto-matic patients believed to have simple cystsalone. The claimed stronger association betweennon-parasitic liver cysts and benign livertumours" is, at first sight, more convincing, butprobably reflects rather the association betweenvon Meyenburg complexes, hepatic fibrosis andpolycystic disease (see below).

POLYCYSTIC DISEASESAdult polycystic disease, of autosomal dominantinheritance with high penetrance, may present atany age with renal manifestations, but althoughhepatic involvement iscommon, clinically signifi-cant liver disease is rare (<15%), and does notusually affect the natural history.26 Hepatic cystsare not often seen before puberty but becomemore frequent with increasing age (<20% inthose under 30,27 28 >75% in those over 7026 27);they are more common in women,28 especiallythose with children. When liver disease occurs itusually accompanies relatively advanced renaldisease (mean creatinine clearance two thirds ofthose without hepatic cysts.28) Polycystic liver

disease may, however, occur in the absence ofsignificant renal involvement. The cysts are linedby biliary type epithelium which retains a func-tional response to secretin.9 Their pathogenesis(akin to that of simple cysts) is considered toreflect failed involution of excess intralobularbile ducts during embryological develop-ment.30 31 Liver biopsy shows portal tract fibrosisand numerous cystically dilated channels; thehepatic parenchyma is otherwise normal but it isnot unusual to find Von Meyenburg complexesin non-cystic parts of the liver (see below). Overtcongenital hepatic fibrosis (see below) certainlyoccurs32 but is unusual.

Apparently spontaneous infection of hepaticcysts is described, usually with gut derivedorganisms, and prompt recognition and drainagegreatly reduces the mortality which may exceed75% with antibiotic therapy alone.33 Mechanicalcholestasis caused by the cysts is rare,3' but acommon topic of case reports: it remains unclearwhy the pressure within the cysts should risesufficiently to cause obstruction, but hormonalresponsiveness29 may be of relevance. Relief ofbiliary obstruction may be satisfactorily pro-vided by percutaneous aspiration.3' In advanceddisease portal hypertension responsible forascites and variceal haemorrhage can becomeproblematic34: good results may be obtainedfrom portosystemic anastomosis. In the majorityof cases, however, management will be that ofchronic renal failure with consideration of renaltransplantation. If there are problems related tothe sheer size of the liver or coexistent sympto-matic portal hypertension, then many units willnow take the opportunity to carry out doubleorgan grafts from the same donor: good resultsare reported, but disappointingly King's was notthe first unit to publish specifically on thistopic.35The autosomal recessive form of polycystic

disease usually presents in infancy with renalfailure and massive kidneys: these patients rarelyhave significant liver disease. Presentations laterin life are associated with less aggressive renaldisease, but with a progressively higher fre-quency of congenital hepatic fibrosis and likeli-hood of portal hypertension.32 Patients with themildest renal involvement are perhaps mostlikely to have an hepatic presentation. Theclinical diagnosis based on renal enlargementand hepatomegaly is supported by biochemicalevidence of renal failure, relatively normalliver function, and the demonstration of renalcysts on utrasound scanning: the hepatic cystsare often too small to be identified sonographi-cally. Histologically, the hepatic cysts appearidentical to those of autosomal dominant cases.

Liver Unit authors were responsible for therecognition of a number of other renal cysticdisorders associated with hepatic fibrosis,36 thelatter apparently indistinguishable from con-genital hepatic fibrosis.

CONGENITAL HEPATIC FIBROSISCongenital hepatic fibrosis deserves its presentinclusion because of the frequency with which itis associated with cystic disease. In its overt formit usually presents in childhood with complica-

S117

Forbes, Murray-Lyon

tions of non-cirrhotic portal hypertension(especially bleeding), but may present in adult-hood;37 asymptomatic forms are encountered atall ages. It is most often associated with therecessive form of polycystic kidney disease, butat least two groups32 38 give convincing accountsof an association between congenital hepaticfibrosis and autosomal dominant polycysticdisease. Equally, in a few cases the kidneys areapparently normal.36 Histologically there is aconsiderable increase in fibrous stroma dis-tributed in broad bands throughout the liverwith focal epithelial components and islands ofnormal lobular architure. Such 'pure' congenitalhepatic fibrosis may be disguised by coexistenthistological features of polycystic disease, andoften by the presence of Von Meyenburg com-plexes. These biliary microhamartomata liewithin dense fibrous stroma and surroundepithelially lined irregular spaces which maycontain bile. The complexes expand within or

adjacent to the portal tracts and, as they areusually multiple, may be confused with micro-metastases. They may be responsible for portalhypertension without more generalised fibrosis.The Liver Unit described the association of

congenital hepatic fibrosis with intra and evenextra hepatic biliary dilatation' producing apicture identical to Caroli's syndrome exceptthat in that situation there is no fibrosis of thecongenital hepatic fibrosis pattern.A recent autopsy study25 which supported the

association of fibrosis with cystic disease addedthe new observation that benign hepaticadenomas are associated with non-parasitic cysts;that the description is ofpolycystic disease ratherthan of simple cysts seems likely, but the signifi-cance of the association is probably of greaterimportance than the semantics of diagnosis.

CAROLI'S SYNDROMECongenital dilatation of the intrahepatic bileducts without obstruction was recognised byCaroli in 1958.39 Pure Caroli's syndrome remainsan unusual diagnosis, with fewer than 150 casesin the world literature. A hereditary component- possibly recessive autosomal transmission - issupported by a report of the condition in twosisters.I The saccular cystic dilatations more

commonly affect the left lobe, and may beassociated with biliary stasis, cholangitis, stoneformation (in about 25%), and pain. Involve-ment of the extra hepatic bile ducts is unusual.Although symptoms may begin in childhood andprogress remorselessly via fibrocholangio-matosis to cirrhosis, a significant proportion ofaffected individuals will remain asymptomatic.The diagnosis should be seriously considered inany child with cholangitis. The combination ofultrasonography and ERCP lead to the diagnosis,although there is a small risk that the latter mayreprecipitate cholangitis. Management dependson correct diagnosis and on the extent of liverinvolvement. Symptomatic disease affectingonly one lobe of the liver can usually be success-

fully resected,4' albeit with considerable diffi-culty if extensive intrahepatic cholelithiasis hasto be dealt with concurrently.42 More extensiveresection, with hepatojejunal anastomosis, is also

possible for multilobular disease,43 and theoption of orthotopic transplantation remains forthe end-stage patient with bilobar disease and/orcirrhosis, so long as sepsis can be controlledpreoperatively.

Caroli's syndrome may be complicated bypancreatitis,45 obstructive portal lymphadeno-pathy4 or extra biliary sepsis,47 and advanceddisease is usually associated with fibrocholangio-matosis; itself often responsible for portal hyper-tension.

Primary sclerosing cholangitis has a number ofcommon features with Caroli's syndrome,especially when dilated duct segments take on a

cystic appearance." Although a firm diagnosis isusually possible there are occasions when the twoconditions cannot be clearly distinguished, and itis possible that an overlap syndrome exists.

Caroli's syndrome is certainly associated withhepatic fibrosis, and a report of Caroli's syn-

drome in renal polycystic disease49 suggests againthat the distinctions between the different cysticdiagnoses are not fundamental but more a matterof degree. Also consistent with this notion, is theimportant observation that Caroli's syndrome isprobably a premalignant condition; frank malig-nancy47 may be preceded by biliary epithelialdysplasia.50

CHOLEDOCHAL CYSTSCholedochal cysts account for approximately 1%of all benign biliary disease5' and as recently as1980 less than 40% were diagnosed preopera-tively.52 They are primarily ofconcern to paedia-tricians and paediatric surgeons as more than60% present before 1Oyearsofage. Thefrequencyis highest in Japan but appears relatively uniformacross the rest of the globe. Numerous largehistorical series have been reported, but thisrelatively common rarity has continued to exertan appeal to authors and editors alike, becauseseveral hundred papers can be identified fromIndex medicus over the past five years alone -

many of these being single case reports. Aconsensus on management is neverthelessemerging despite the absence ofcontrolled trials.Almost all authorities now use modifications ofthe Alonso-Lej classification,53 typically dividingcholedochal cysts into four groups (Fig 1). Thecommonest (type I) comprises a fusiform or

saccular dilatation ofthe extrahepatic biliary treewith a normal or stenotic distal common bileduct. Less often cysts appear as common bile

Figure 1: Classification ofcholedochal cysts (afterAlonso-Lejet al.)"

I II III IV

S118

Cystic disease ofthe liver and biliary tract

bb

Figure 2: Illustration oflong common pancreatobiliarychannel. Left: typicaljunction ofduct within sphincterzone. Right: long common channel with confluence ofpancreatic and biliary systems proximal to sphincter.a - ampulla; b - bile duct; p - pancreatic duct.

duct diverticula (type II), are contained withinthe intraduodenal course of the duct (type III),or are more generalised, involving both intra andextrahepatic duct systems (type IV) (Fig 1).There is evidently a point at which it becomesdifficult to distinguish between intrahepaticcholedochal cyst and Caroli's syndrome; reportsclaiming heterogeneity of Caroli's syndrome54apparently fail to appreciate the semantic natureof diagnosis in this situation. It is probablyreasonable to consider choledochal cyst a part ofthe continuum already discussed in respect ofthepurely intrahepatic cystic conditions.

It is almost certain that choledochal cysts areof congenital origin, but the pathogenesis is lessclear. It has been speculated that there is adifferential abnormality in biliary epithelial pro-liferation in the embryo such that the solid coreof proto-bile duct takes on a cystic form55; thereis, however, no good evidence to support this. Itis probable that reflux ofpancreatic juice into thebiliary tree is of relevance, however.56 A verylarge proportion of (and possibly all) patientswith choledochal cyst have an unusually longcommon channel between the junction of thecommon bile duct and the main pancreatic ductand their joint outflow into the duodenum(Fig 2).57 The increased intraductal pressureassociated with this anomaly predisposes toreflux and may also be associated with ectasia ofthe common channel57 (or indeed with pan-creatitis.)58 High concentrations of pancreaticenzymes within cyst fluid are common andpancreatitis may be diagnosed erroneously whenhyperamylasaemia also occurs.59 A recent reportofcholedochal cyst in mother and daughter, bothof whom had a long common channel and thesame type of cystO suggests that a hereditaryelement is important in some cases. Concurrenthepatic fibrosis has not attracted much attention,but certainly occurs,37 and lends support to theinclusion of choledochal cystic disease in thefibrocystic continuum.

Choledochal cysts may remain asymptomaticfor many years (and possibly for life) but themajority will at some time cause one or more ofthe classical triad of pain, obstruction and pal-pable mass, often in association with cholangitis,or exhibit one of the less common featuresenumerated by the Department of Surgery atKing's,61 such as ascites, variceal haemorrhage,or biliary peritonitis from spontaneous rupture.

Many patients will initially respond to simplemedical measures, allowing time for carefulassessment. Continued conservative therapy,provision of internal (or very occasionallyexternal) drainage, or excision of the cyst mustthen be considered. The anatomy will usually bereadily apparent from ultrasound scanning andERCP, but will in a few cases remain enigmaticdespite the full range of currently availableradiological techniques. Treatment optionsshould be considered in the light of the collectedhistorical series. It appears that about 50% ofpatients having drainage procedures will needreoperation - usually because of recurrent sepsisor cholestasis,62 and one author63 makes thestartling claim that his study 'provides conclusiveevidence that any anastomosis of cyst wall to thegastrointestinal tract will ultimately result instricture and cholangitis'. Moreover, at least onegroup'M have found (admittedly without takingfull account of the technical considerations thatdetermined the original operation), that theperioperative mortality is substantially higherfor drainage procedures than for excision (8-6%v 0%). Excisions may vary between a chole-cystectomy like removal of a type II cyst to avirtual extirpation of the biliary tree; mostauthors now prefer hepatojejunostomy.Excisions are not without their complications,however. One of the most experienced groups62recently described a reoperation rate of nearly10% in 73 excisions - although all seven ulti-mately did well. Major revisions were needed inmost cases - usually conversion from hepatoduo-denostomy to hepato-Roux-Y-jejunostomy afterearly haemorrhage/leak or late stenosis of theoriginal anastomosis.

Carcinoma (adenocarcinoma in more than90%) is now well recognised as a complication ofcholedochal cysts. The life time risk for theindividual may be as high as 50% if non-resectional surgery has been carried out.65 Bilestagnation, continual pancreatic reflux, andrelated chronic ulceration and regeneration ofthe cyst epithelium are thought causallyrelevant.'M Investigation of an 18 year old patientwith a choledochal cyst showed an increase insecondary bile salts (most of which were uncon-jugated) within the cyst fluid, despite an entirelynormal bile salt pattern in intrahepatic bile. Theauthors attribute this to bacterial colonisationwithin the cyst, and speculate that this causedthe patient's biliary epithelial metaplasia andcould contribute to the development of neo-plasia.67 It is probable that the cancer risk issubstantially higher in patients presenting forthe first time as adults, and attributed risks ofbetween 5 and 40% have been suggested.66Review of the computed tomography files in oneJapanese institution68 revealed that eight of 35consecutive adult cases of choledochal cyst hadcomplicating malignancy. It must be noted thatthe studies with the longest follow up periodshave consistently given the highest figures.Although the most common site of origin formalignant change is the posterior cyst wall, it isevident that there is a biliary 'field' defect asprimary carcinoma may occur in biliary tissueuninvolved in the cyst, a striking 46% occurringin apparently normal gall bladder in one series.'

S119

Forbes, Murray-Lyon

Prognosis once carcinoma has occurred is uni-formly poor.

It is reasonable to conclude that choledochalcysts should always be excised with as completeresection of the (extrahepatic) biliary tree as

possible. Such a policy should minimise post-operative cholangitis/cholestasis and achieve a

low frequency ofreoperation; it should also, withluck, eliminate the risk of malignant transforma-tion within retained cystic components, or

apparently normal more proximal biliaryepithelium. When comprehensive excisionimplies a major pancreatic resection as well as

dissection up into the liver, however, dogmaticviews may have to be tempered, and especially so

if the patient is asymptomatic!

CYSTIC NEOPLASIAHepatic cysts are not often neoplastic, but cystictumours are recognised occasionally. Multi-locular cystadenoma occurs almost exclusively inmiddle aged women who typically present with a

rapidly growing upper abdominal mass; aboutone third develop clinical cholestasis.69 Thedistinction from simple cysts is straightforwardhistologically but less obvious clinically or radio-logically: the presence of multiple cystic areas

(usually several small lesions around a largercyst), the presence ofbiliary obstruction, and therate of progression are useful pointers. Surgicalresection is advised as malignant transformationmay occur.Myxomatous tumours ofthe liver are evidently

very rare, but may be responsible for a painfulmass70: the single reported case was probablybenign. Benign cystic hepatoblastoma occurs inpreschool children,7' usually presenting as anabdominal mass with multiocular cysts on in-vestigation; transformation to cystadenocar-cinoma may occur if unresected. Malignantcystic hepatoblastoma is a distinct and very rare

paediatric tumour where the malignant com-

ponent of the tumour is stromal rather thanepithelial,71 the prognosis is poor.

Cystic metastases from carcinoma of the ovaryand kidney, and from intra-abdominal sarcomasare also described. The malignant nature oftheselesions is usually apparent from associatedparenchymal disease, or from the classical sono-

graphic appearance of a malignant cyst, withthick irregular walls studded by nodules.' 2 Thediagnosis will usually be clarified by aspirationcytology, but serial scanning is almost as helpfulas cystic metastases typically enlarge rapidly.

OTHER CYSTIC DISORDERSVarious other cystic disorders have been des-cribed, of which 'ectopic' pancreatic pseudocystis perhaps the most common, but accuratediagnosis is otherwise unusual without histo-logical examination of material obtained atsurgery or autopsy. Any list is unlikely to becomprehensive.

Occlusion of small portal vein branches fromany cause (particularly compression by adjacenthepatic tumour) can produce areas of apparentinfarction manifest as small cystic areas; as tissuenecrosis is absent these are now generally known

as pseudoinfarcts of Zahn72 and have no prog-nostic significance in their own right.

Peliosis hepatis was originally described in theterminal cachexia of disseminated tuberculosis,but is now most often seen in women oncombined contraceptive pill. Work from thisunit clarified a link between peliosis and longterm androgen therapy,73 and associations withother conditions such as HIV infection are alsorecognised.74 It is probably underdiagnosed asinvestigation is embarked upon only in thosewho develop hepatomegaly or abnormalities ofliver function. Histologically there are bloodfilled cavernous cysts in continuity with thesinusoids. Cysts vary in size but may reach 5 mmin diameter and allow sonographic recognition.Peliosis is usually harmless, and it will oftenregress ifa predisposing factor can be withdrawn.Rarely, liver failure or hepatic rupture mayoccur. Sinusoidal ectasia appears to be a milderform of peliosis and may represent its earliestmanifestations.

Less than 20 cases of ciliated hepatic foregutcyst have been described.75 They are thoughtanalogous to bronchogenic cysts, with whichthey share a ciliated mucus secreting pseudo-stratified columnocuboidal lining with amuscular wall and surrounding fibrous tissue.Characteristically they have a subcapsular siteand are usually less than 3 cm in diameter.The sonographic findings include heterogeneousechogenic semi fluid contents (which may at firstsight appear solid) within a well demarcatedlesion, and aspiration reveals viscous mucinousexudative fluid. They are non-neoplastic andrequire no treatment if the diagnosis is certain.

Endometriosis may also be responsible forhepatic cysts, successfully treated in the firstreport76 by resection. Had the diagnosis beenrealised preoperatively might hormonal therapyhave sufficed?

Post traumatic hepatic cysts are recognised2 inaddition to the cystic spaces that may remainafter successful non-drainage treatment for liverabscess. A similarphenomenon has been reportedfrom the Liver Unit after apparently curativehepatic artery embolisation (for metastaticapudoma),77 and is probably not uncommonafter this intervention.

Cystic fibrosis is not, it appears, associatedwith higher frequencies of hepatobiliary cyststhan would be expected by chance.

ConclusionThere is much in the collected literature onhepatic and biliary cystic disorders to suggestcommon threads in aetiology, pathogenesis,clinical presentation, complications and manage-ment (assuming that cysts of infective or neo-plastic origin are excluded). The strong associa-tion of adult polycystic kidney disease withpolycystic liver disease is not complete, and itneed not be assumed that the hepatic cysts areencoded for by the same gene defect. The variousoverlap syndromes earlier alluded to allow thepossibility that even if polycystic liver disease isalso of autosomal dominant inheritance, it hasvariable expression with relatively low pene-trance. Almost all hepatobiliary cysts are lined

S120

Cystic disease ofthe liverand biliary tract S121

by biliary epithelium or epithelium with strongbiliary characteristics. All most probablydevelop in the prenatal period (and clinicalpresentation is generally commoner in child-hood). All are more common in women and thereare suggestions that hormonal factors may berelevant. All share an assocation with hepaticfibrosis, although the strength of the associationadmittedly varies according to the major charac-teristics of the cystic disorder. Most, and prob-ably all, of the cysts engender a substantiallyincreased risk of biliary tract malignancy (andperhaps hepatic neoplasia also), the added riskappearing to be greatest where the cyst is incommunication with the bilary tree. It is unusualfor cysts themselves to be responsible for symp-toms, but the complications, including those ofportal hypertension, that lead to presentation arecommon to all varieties. Management canusually afford to be expectant except for chole-dochal cysts where, because of the cancer risk,elective resection is indicated. There seems everyreason to support and extend the belief8 thatthere is a family of interlocked and overlappinghepatobiliary fibrocystic disorders - a view heldin part for many years on the Liver Unit.79

1 Wilson MA, Ruzicka FF, eds. Modern imaging ofthe liver. NewYork: Marcel Dekker, 1989.

2 Gaines PA, Sampson MA. Prevalence and characterization ofsimple hepatic cysts by ultrasound examination. BrJ Radiol1989; 62: 335-7.

3 Wilcox DM, Weinreb JC, Lesh P. MR imaging of a hemor-rhagic hepatic cyst in a patient with polycystic liver disease.J7 ComputAssist Tomogr 1985; 9: 183-5.

4 Wittenberg J, Stark DD, Forman BH, et al. Differentiation ofhepatic metastases from hepatic hemangiomas and cysts byusing MR imaging. AmJ Roentgenol 1988; 151: 79-84.

5 McCorkell SJ. Unintended percutaneous aspiration of pul-monary echinococcal cysts. Am J Roentgenol 1984; 143:123-6.

6 Mueller PR, Dawson SL, Ferrucci TJ Jr, Nardi GL. Hepaticechinococcal cyst: successful percutaneous drainage.Radiology 1985; 155: 627-8.

7 Filice C, Pirola F, Brunetti E, Dugheti S, Strosselli M,Foglieni CS. A new therapeutic approach for hydatid livercysts: aspiration and alcohol injection under sonographicguidance. Gastroenterology 1990; 98: 1366-8.

8 Murphy BJ, Casillas J, Ros PR, Morillo G, Albores-SaavedraJ, Rolfes DB. CT appearance of cystic masses of the liver.Radiographics 1989; 9: 307-22.

9 Williamson RCN, Ramus NI, Shorey BZ. Congenital solitarycysts of the liver and spleen. BrJ Surg 1978; 65: 871-6.

10 Pinto MM, Kaye AD. Fine needle aspiration of cystic liverlesions. Cytological examination and carcinoembryonicantigen assay of cyst contents. Acta Cytol 1989; 33: 852-6.

11 Roisman I, Barak V, Fields S, Bloom R, Okon E, Manny J,Solitary non-parasitic cyst of the liver: a rare cause ofabdominal distension. AmJ Gastroenterol 1989; 84: 1095-9.

12 Clinkscales NB, Trigg LP, Poklepovic J. Obstructive jaundicesecondary to benign hepatic cyst. Radiology 1985; 154:643-4.

13 Ayyash K, Haddad J. Spontaneous rupture of a solitarynonparasitic cyst of the liver. Acta Chir Scand 1988; 154:241-3.

14 Rashed A, May RE, Williamson RCN. The management oflarge congenital liver cysts. Postgrad Med J 1982; 58:536-41.

15 Lotz GW, Stahlschmidt M. Intra-abdominal bleeding afterrupture of hepatic cyst. South MedJ 1989; 82: 667.

16 Akriviadis EA, Steindel H, Ralls P, Redeker AG. Spontaneousrupture of nonparasitic cyst of the liver. Gastroenterology1989; 97: 213-5.

17 Chang D. Solitary non-parasitic liver cyst treated by cyst-gastrostomy. PostgradMedJr 1988; 64: 459-61.

18 Trinkl W, Sassaris M, Hunter FM. Nonsurgical treatment ofsymptomatic nonparasitic liver cyst. Am J Gastroenterol1985; 80:907-11.

19 Tsuboi Y, Yamada H, Arai T, et al. Treatment of large hepaticcyst by absolute ethanol injection into the cavity. JikeikaiMedJt 1987; 34: 583-92.

20 Huguier M, Paquet JC, Roland J, Houry 5, Lacaine F. Biliarycysts of the liver. DigSurg 1990; 7: 93-6.

21 Hayashi I, Tomoda H, Tanimoto M, et al. Mucoepidermoidcarcinoma arising from a preexisting cyst of the liver. J SurgOncol 1987; 36: 122-5.

22 Rehulova E, Dite P. Multifocal carcinoma arising from acongenital cyst of the liver and the kidneys. Cesk Patol 1981;17: 198-203.

23 Lynch MJ, McLeod MK, Weatherbee L, GilsdorfJR, GuiceKS, Eckhauser FE. Squamous cell cancer of the liver arising

from a solitary benign nonparasitic hepatic cyst.AmJ Gastroenterol 1988; 83: 426-31.

24 Devine P, Ucci AA. Biliary cystadenocarcinoma arising in acongenital cyst. Hum Pathol 1985; 16: 92-4.

25 Karhunen PJ, Penttila A, Liesto K, Mannikko A, MottonenM. Benign bile duct tumours, non-parasitic liver cysts andliver damage in males. J Hepatol 1986; 2: 89-99.

26 Grunfeld J-P, Albouze G, Jungers P, et al. Liver changes andcomplications in adult polycystic kidney disease. AdvNephrol 1985; 14: 1-20.

27 Dalgaard OZ. Bilateral polycystic disease of the kidneys: afollow-up of284 patients and their families. Acta Med Scand(suppl) 1957; 328: 1-255.

28 Gabow PA, Johnson AM, Kaehny WD, Manco-Johnson ML,Duley IT, Everson GT. Risk factors for the development ofhepatic cysts in autosomal dominant polycystic kidneydisease. Hepatology 1990; 11: 1033-7.

29 Everson GT, Emmett M, Brown WR, Redmond P, ThickmanD. Functional similarities of hepatic cysts and biliaryepithelium: studies offluid constituents and in vivo secretionin response to secretin. Hepatology 1990; 11: 557-65.

30 Moschowitz E. Non-parasitic cysts (congenital) of the liver,with a study of aberrant bile ducts. AmJ Med Sci 1906; 131:674-99.

31 Van Erpecum KJ, Janssens AR, Terpstra JL, Tjon-A-ThamRTO. Highly symptomatic adult polycystic disease of theliver: a report of 15 cases. J Hepatol 1987; 5: 109-17.

32 Cobben JM, Breuning MH, Schoots C, Kate LPT, Zerres K.Congenital hepatic fibrosis in autosomal-dominant poly-cystic kidney disease. Kidney Int 1990; 38: 880-5.

33 Telenti A, Torres VE, Gross JB Jr, Van Scoy RE, Brown ML,Hattery RR. Hepatic cyst infection in autosomal dominantpolycystic kidney disease. Mayo Clin Proc 1990; 65: 933-42.

34 Ratcliffe PJ, Reeders S, Theaker JM. Bleeding oesophagealvarices and hepatic dysfunction in adult polycystic kidneydisease. BMJ 1984; 288: 1330-1.

35 Starzl TE, Reyes J, Tzakis A, Mieles L, Todo S, Gordon R.Liver transplantation of polycystic liver disease. Arch Surg1990; 125: 575-7.

36 Murray-Lyon IM, Shilkin KB, Laws JW, Illing RC, WilliamsR. Non-obstructive dilatation of the intrahepatic biliary treewith cholangitis. QJ Med 1972; XLI: 477-89.

37 De Vos M, Barbier F, Cuvelier C. Congenital hepatic fibrosis.JHepatol 1988; 6: 222-8.

38 Tazelaar HD, Payne JA, Patel NS. Congenital hepaticfibrosis and asymptomatic familial adult-type polycystickidney disease in a 19-year-old woman. Gastroenterology1984; 86:757-60.

39 Caroli J, Soupault R, Kossakowski J, et al. La dilatationpolykystique congenitale des voies biliaires intrahepatiques.Essai de classification. Sem HOp Paris 1958; 34: 488-95.

40 Hoglund M, Muren C, Schmidt D. Caroli's disease in twosisters: diagnosis by ultrasonography and computed tomo-graphy. Acta Radiol 1989; 30: 459-62.

41 Borda F, Uribarrena R, Puras A, Ortiz H, Rivero-Puente A.Resection hepatique pour maladie de Caroli limitee au lobegauche. Gastroenterol ClinBiol 1982; 6: 511-2.

42 Mercardier M, Chigot JP, Clot JP, Langlois P, Lansiaux P.Caroli's disease. WorldJSurg 1984; 8: 22-9.

43 Aeberhard P. Surgical management of Caroli's disease involv-ing both lobes of the liver. BrJ Surg 1985; 72: 651-2.

44 Barker EM, Kallideen JM. Caroli's disease: successful man-agement using permanent-access hepaticojejunostomy.BrJSurg 1985; 72: 641-3.

45 Boyle MJ, Doyle GD, McNulty JG. Monolobar Caroli'sdisease. AmJ Gastroenterol 1989; 84: 1437-44.

46 Chaimoff C, Lurie B, Gal R. Budd-Chiari-like syndrome inCaroli's disease. J Clin Gastroenterol 1986; 8: 586-8.

47 Rogstad K, Freeman J, Moorghen M, Record CO. Difficultyin diagnosing complications of Caroli's disease. J ClinGastroenterol 1986; 8: 582-5.

48 Geneve J, Dubuc N, Mathieu D, Zafrani ES, Dhumeaux D,Metreau J-M. Cystic dilatation of intrahepatic bile ducts inprimary sclerosing cholangitis.JHepatol 1990; 11: 196-9.

49 Jordon D, Harpaz N, Thung SN. Caroli's disease and adultpolycystic kidney disease: a rarely recognised association.Liver 1989; 9: 30-5.

50 Fozard JBJ, Wyatt JI, Hall RI. Epithelial dysplasia in Caroli'sdisease. Gut 1989; 30: 1150-3.

51 Saxena R, Pradeep R, Chander J, et al. Benign disease of thecommon bile duct. BrJ Surg 1988; 75: 803-6.

52 Yamaguchi M. Congenital choledochal cyst. Analysis of 1433patients in the Japanese literature. Am J Surg 1980; 140:653-7.

53 Alonso-Lej F, Rever WB, Pessagno DJ. Congenital chole-dochal cyst, with a report of 2, and an analysis of 94, cases.IntAbstrSurg 1959; 108: 1-30.

54 Tandon RK, Grewal H, Anand AC, Vashisht S. Caroli'ssyndrome: a heterogeneous entity. AmJ Gastroenterol 1990;85: 170-3.

55 Olbourne NA. Choledochal cysts - a review of the cysticanomalies of the biliary tree. Ann R Coll Surg Engl 1975; 56:26-32.

56 Vanderpool D, Lane BW, Winter JW, Ettinger J. Choledochalcysts. Surg Gynecol Obstet 1988; 167: 447-51.

57 Wiedmeyer D, Stewart ET, Dodds WJ, Geenen JE, VennesJA, Taylor AJ. Choledochal cyst: findings on cholangio-pancreatography with emphasis on ectasia of the commonchannel. AmJ Roentgenol 1989; 153: 969-72.

58 Forbes A, Leung JWC, Cotton PB. Relapsing acute andchronic pancreatitis. Arch Dis Child 1984; 59: 927-34.

59 Todani T, Urushihara N, Watanabe Y, et al. Pseudopan-creatitis in choledochal cyst in children: intraoperative studyof amylase levels in the serum. J Pediatr Surg 1990; 25:303-6.

S122 Forbes, Murray-Lyon

60 Iwafuchi M, Ohsawa Y, Naito M, Maruta Y, Saito H. Familialoccurrence of congenital bile duct dilatation. J Pediatr Surg1990; 25: 353-55.

61 Tan KC, Howard ER. Choledochal cyst: a 14 year surgicalexperience with 30 patients. BrJ Surg 1988; 75: 892-5.

62 Todani T, Watanabe Y, Toki A, Urushihara N, Sato Y.Reoperation for congenital choledochal cyst. Ann Surg 1988;207: 142-7.

63 Joseph VT. Surgical techniques and long-term results in thetreatment of choledochal cyst. J Pediatr Surg 1990; 25:782-7.

64 Saing H, Tam PKH, Lee JMH, Pe-Nyun. Surgical manage-ment of choledochal cysts: a review of 60 cases. J PediatrSurg 1985; 20: 443-8.

65 Flanigan DP. Biliary carcinoma associated with biliary cysts.Cancer 1977; 40: 880-3.

66 Todani T, Watanabe Y, Toki A, Urushihara N. Carcinomarelated to choledochal cysts with internal drainage opera-tions. Surg Gynecol Obstet 1987; 164: 61-4.

67 Reveille RM, Van Stiegmann G, Everson GT. Increasedsecondary bile acids in a choledochal cyst: possible role inbiliary metaplasia and carcinoma. Gastroenterology 1990; 99:525-7.

68 Yoshida H, Itai Y, Minami M, Kokubo T, Ohtomo K, KurodaA. Biliary malignancies occurring in choledochal cysts.Radiology 1989; 173: 389-92.

69 Marsh JL, Dahms B, Longmire WP Jr. Cystadenoma andcystadenocarcinoma of the biliary system. Arch Surg 1974;109:41-3.

70 Edmondson HA. Tumors of the liver and intrahepatic bileducts. Atlas of tumor pathology, section VII, fascicle 23.Washington DC: Armed Forces Institute of Pathology,1974.

71 Landing BH. Tumors of the liver in childhood. In: Okuda K,Peters RL, eds. Hepatocellular Carcinoma. New York: WileyMedical, 1976.

72 Holdstock G, Millward-Sadier GH, Wright R. Hepaticchanges in systemic disease. In: Wright R, Millward-SadlerGH, Alberti KGMM, Karran S, eds. Liver and biliarydisease, London: Baillere Tindall, 1985.

73 Paradinas FJ, Bull TB, Westaby D, Murray-Lyon IM.Hyperblasia and prolapse of hepatocytes into hepatic veinsduring longterm methyltestosterone therapy: possiblerelationships of these changes to the development of peliosishepatitis and liver tumours. Histopathology 1977; 1: 225-46.

74 Perkocha LA, Geaghan SM, Yen BTS, et al. Clinical andpathological features of bacillary peliosis hepatis in associa-tion with human immuno deficiency virus infection.NEnglJMed 1990; 323: 1581-6.

75 Kadoya M, Matsui 0, Nakanuma Y, et al. Ciliated hepaticforegut cyst: radiologic features. Radiology 1990; 175:475-7.

76 Finkel L, Marchevsky A, Cohen B. Endometrial cyst of theliver. AmJ Gastroenterol 1986; 81: 576-8.

77 Lee S-M, Forbes A, Williams R. Metastatic islet cell tumoursecreting insulin and glucagon: successful treatment byhepatic artery embolization and chemotherapy. EurJ SurgOncol 1988; 14: 265-8.

78 Summerfield JA, Nagafuchi Y, Sherlock S, Cadafalch J,Scheuer PJ. Hepatobiliary fibropolycystic diseases: a clinicaland histological review of 51 patients. J Hepatol 1986; 2:141-56.

79 Murray-Lyon IM, Ockenden BG, Williams R. Congenitalhepatic fibrosis - is it a single clinical entity? Gastroenterology1973;64:653-6.