Primary sclerosing cholangitis

Overview

Primary sclerosing cholangitis (PSC) is a chronic, idiopathic cholangiopathy characterized histologically by peribiliary inflammation and fibrosis. PSC can progress to cirrhosis and is a risk factor for hepatobiliary and colonic carcinogenesis. The median liver transplant (LT) free survival time is approximately 15 years. Despite ongoing research over the past several decades, the etiopathogenesis of PSC remains poorly understood; as a result, effective pharmacologic therapy for PSC has not been established. ^, Despite being a relatively rare disorder, PSC is the fifth most common indication for LT in the United States (US) (UNOS) and the leading indication for LT in several other countries ^, (see Chapter 105 ). LT is presently the only proven life-extending therapy for patients with end-stage PSC; although LT is potentially curative, it is offered to only selected patients, and even suitable candidates can experience recurrent PSC or hepatobiliary malignancy post-LT. ^,

Epidemiology

PSC more commonly affects men than women (2:1) and has a peak incidence in the fourth decade of life. ^, The age-adjusted incidence of PSC in the US is 1.25 per 100,000 for Caucasian men and 0.54 per 100,000 for Caucasian women. Recent data suggest the prevalence of PSC has been rising in both sexes, which may be attributable to an increase in incidence and not decrease in deaths. The calculated prevalence of PSC in the US has been reported as 20.9 and 6.3 per 100,000 people, respectively, corresponding to approximately 30,000 cases nationwide. Similar statistics on PSC prevalence have been reported in Canada, Northern Europe, ^, and New Zealand. Limited available data suggest a similar prevalence for African American patients in the US compared with European American patients, but African American patients tend to be younger and have a higher model for end-stage liver disease (MELD) at the time of LT listing. Epidemiologic data regarding the geographic distribution of PSC remains poorly defined in most other countries.

PSC is strongly associated with inflammatory bowel disease (IBD). Approximately 70% of Western (e.g., US, UK) patients with PSC are co-diagnosed with IBD, most commonly ulcerative colitis (UC) ^, ; conversely, only approximately 3% to 5% of patients with IBD have PSC. It is worth noting, though, that in far Eastern cohorts, only approximately one-third of patients with PSC have concomitant IBD, a discrepancy that may be attributable to genetic and/or environmental factors.

Clinical presentation

Up to 40% of patients with PSC are asymptomatic at the time of diagnosis and incidentally found to have elevated serum liver enzymes, most commonly alkaline phosphatase (ALP) and alanine aminotransferase (ALT), which leads to further evaluation and ultimately the diagnosis of PSC. ^, When present, the most common symptoms at the time of PSC diagnosis include abdominal pain, pruritus, diarrhea, jaundice, fatigue, and fever. The two most frequently encountered physical examination signs are hepatomegaly and splenomegaly, although the physical examination is often unrevealing. On occasion, a patient may present with manifestations of advanced cirrhosis and portal hypertension, including ascites or gastrointestinal bleeding secondary to varices. Children with PSC relatively frequently have liver disease with features of autoimmune hepatitis (AIH) as discussed later (see “ Associated Diseases ” section). ^, Therefore the clinical presentation of PSC varies considerably, depending in part on the disease stage at the time of diagnosis and the age at diagnosis.

Diagnosis

The most common biochemical abnormality in patients with PSC is an elevated ALP, reflecting chronic cholestasis, and may be the only laboratory abnormality identified. Although some patients may have a normal ALP, it is typically about twofold to threefold higher than the upper limit of normal. Unlike other autoimmune liver diseases, such as autoimmune hepatitis or primary biliary cirrhosis, there is no serologic test that aids in diagnosis but often just nonspecific elevated autoantibodies.

Noninvasive imaging, in particular magnetic resonance cholangiopancreatography (MRCP), can be used to diagnose PSC (see Chapter 16 ). Cholangiographic features of PSC include multifocal intra and/or extrahepatic biliary strictures, proximal segmental ductal dilatation, and a “pruned” biliary tree ( Fig. 41.1 ). Given advances in noninvasive imaging techniques, including MRCP, endoscopic retrograde cholangiopancreatography (ERCP) is generally no longer needed to make a diagnosis of PSC. Instead, ERCP is now typically reserved for evaluation of unexplained hepatobiliary symptoms, biliary obstruction caused by a suspected dominant stricture, or possible/known malignancy, among other indications ( Fig. 41.2 ; see Chapters 20 and 30 ).

A liver biopsy is not required for the diagnosis of PSC, especially if the patient fits the classical pattern, including a patient with underlying IBD, elevated ALP, and typical findings on MRCP. The major role of liver biopsy in PSC, if performed, is to: (1) exclude other or co-existing causes of liver disease (e.g., autoimmune hepatitis), (2) diagnose small-duct PSC, and (3) determine the disease stage. Pathologic features compatible with PSC include chronic cholangitis, ductular proliferation, and periductal fibrosis ( Fig. 41.3 ) but can vary depending on disease stage. ^{, ,}

It is important to distinguish between PSC and other causes of liver disease, especially secondary sclerosing cholangitis ( Table 41.1 ), because secondary sclerosing cholangitis generally originates from known pathologic processes (e.g., biliary trauma, malignancy, infection) and may respond well to pharmacotherapy. An important example of secondary sclerosing cholangitis is IgG4-associated cholangiopathy, which is part of a spectrum of systemic fibroinflammatory disorders (IgG4-related diseases) that can affect multiple organ systems, are characterized by elevated serum and/or tissue IgG4 levels, and generally respond to corticosteroid therapy. Novel qPCR quantifying the IgG4/IgG RNA ratio in blood and next-generation sequencing showing dominant IgG4+ B-cell receptor clones in patients with IgG4-related disease are both new tests that can aid in differentiating patients with IgG4-realted disease from those with PSC ^, (see Chapter 43 ).

Other serologic abnormalities

Although the classic serum biochemical pattern in PSC is cholestatic, other or concomitant abnormalities may also be seen. For instance, serum aminotransferase levels may also be increased in many patients, albeit only modestly (to less than three times the upper limit of normal). Those with markedly elevated aminotransferase levels may show concomitant serologic and histologic features of AIH, thus suggesting PSC-AIH overlap syndrome, as discussed further in a subsequent section (see “ Associated Diseases ” section). Because patients with PSC may develop an overlap syndrome years after the initial diagnosis of PSC, periodic monitoring of aminotransferases (together with ALP and bilirubin) is advisable.

Serum bilirubin levels are normal in 60% of patients at diagnosis but tend to rise as PSC progresses. An abrupt, sustained increase in conjugated bilirubin may indicate the presence of a dominant biliary stricture, bile duct stone, or the development of cholangiocarcinoma (CCA) ; therefore it should prompt additional investigation (e.g., MRCP). Serum copper and ceruloplasmin and hepatic and urinary copper levels are often abnormal. Hepatic copper levels can be increased to the degree seen in Wilson’s disease and primary biliary cirrhosis (PBC) and is a reflection of prolonged cholestasis.

Elevated serum IgM, IgE, IgG, and total IgA has been reported in approximately 45%, 40%, 25%, and 10% of patients with PSC, respectively. ^{, , ,} Patients with an increase in one immunoglobulin isotype will generally have increased levels of other isotypes as well. Although the relevance of hyperglobulinemias in PSC remains uncertain, it is conceivable that they may have as yet unrecognized prognostic and/or pathophysiologic implications (e.g., as with IgG4-related cholangiopathy). Auto-antibody testing in PSC is of limited value.

Imaging modalities

Cholangiography is necessary for diagnosing large-duct PSC. MRI/MRCP is a noninvasive modality that visualizes the biliary tree and has largely replaced the prior gold standard invasive technique of ERCP (see Chapters 16 , 20 , and 30 ). The International PSC Study Group has recommended MRCP as the initial diagnostic imaging technique in patients suspected of having PSC. Classically, both the intrahepatic and extrahepatic biliary tree is involved, but variations of PSC exist ( Table 41.2 ), and involvement may become more extensive over time in those presenting with only intra or only extrahepatic disease. In addition to location, strictures can also vary in length and severity (i.e., degree of fibrosis and obstruction). Some strictures may also harbour malignancy, and indeed distinguishing between benign and malignant dominant strictures in PSC represents a major ongoing challenge, ^, as discussed later (see “ Cholangiocarcinoma ” section).

The emergence of MRCP has been partly driven by ERCP being an invasive procedure requiring anaesthesia and carrying a considerable risk of adverse events, such as acute pancreatitis, bleeding, perforation, and acute cholangitis. Patients with PSC, in particular, have a higher incidence of acute cholangitis compared with patients without PSC undergoing ERCP, whereas the risk of other adverse events appears to be similar. Because of the risks of ERCP, MRCP has emerged as a noninvasive substitute for ERCP (see Fig. 41.2 ). MRCP has diagnostic accuracy comparable with that of ERCP and results in cost savings when used as the initial diagnostic strategy. ^, MRCP also lends itself well to simultaneous MR elastography of the liver, which is a technique used to quantify hepatic fibrosis (expressed in kilopascals) and distinguish patients with and without cirrhosis.

Abdominal ultrasound and computed tomography (CT) are useful in monitoring patients and evaluating for biliary stones, hepatobiliary malignancy, and other disease complications once a diagnosis of PSC has already been made (see Chapter 16 ). Percutaneous cholangiography (PTC) allows access to the biliary tree when ERCP is not technically feasible (see Chapters 20 and 31 ). Positron emission tomography (PET)/CT has a high sensitivity for CCA in a dominant stricture and has a role in combination with brush cytology for monitoring PSC patients who have a dominant stricture.

Histopathology

Classic histologic findings on liver biopsy include paucicellular, mixed, nonsuppurative portal tract inflammation, cholangitis (with or without ductopenia), cholestasis, and a periductal cuff of fibrous sheets and edema that acquires a classic “onion skin” appearance (see Fig. 41.3 ). This classic finding is nearly pathognomonic but is seen in fewer than 10% of PSC liver biopsies (but more frequently in larger surgical specimens).

It is important to recognize and distinguish fibro-obliterative cholangitis of PSC from alternative etiologies including: (1) PBC, (2) mechanical obstruction of larger bile ducts, (3) ductopenic rejection after LT, (4) cholangiopathy of acquired immunodeficiency syndrome (AIDS), and (5) patients having undergone hepatic arterial infusion of floxuridine (FUDR). Distinguishing PSC from PBC, for example, can be based on the involvement of extrahepatic and large intrahepatic bile ducts in the former, as well as the milder, mixed inflammatory infiltrate, but this distinction may be difficult to make in some cases without appropriate clinical correlation. Moreover, granulomata, thought to be a feature of PBC, may be seen in approximately 4% of liver biopsies from patients with PSC (thus further emphasizing the importance of additional clinical data.

There are three main liver histology grading systems relevant to PSC: Nakanuma, Ishak, and Ludwig. The Ludwig system has been the most commonly used and is based on the degree of extension of fibroinflammatory changes in the hepatic parenchyma, ranging from stage 1 (portal inflammation) to stage 4 (biliary cirrhosis; Table 41.3 ). A head-to-head comparison of the three scoring systems has shown that all are independent predictors of long-term outcomes in patients with PSC but that the Nakanuma has the highest predictive value.

Etiopathogenesis

Although the pathogenesis of PSC remains unclear, it is generally accepted that environmental exposures and genetic underpinnings drive inflammation and fibrosis in PSC via alterations in the gut microbiota, autoimmunity, bile acid composition, and a proinflammatory cholangiocyte phenotype. Cholangiocytes, the epithelial cells lining the bile ducts, are a morphologically, biochemically, and functionally heterogeneous and highly dynamic population of cells that are not only a target of injury in PSC but may also be directly and actively involved in its pathogenesis. ^, For example, in response to recognition of pathogen associated molecular patterns (PAMPs) and other stimuli such as infection or ischemia, cholangiocytes are activated to express a number of proinflammatory cytokines (e.g., tumor necrosis factor [TNF]-α, interleukin [IL]-6, IL-8) and other bioactive molecules. Biosynthesis and secretion of these signaling mediators makes up part of the biliary innate immune and repair response and mediates recruitment and activation of T cells, macrophages, neutrophils, natural killer cells, and other resident and recruited cells in the peribiliary environment. ^{, ,}

Several hypotheses regarding the etiopathogenesis of PSC have been proposed, ^, of which two are described herein, and both of which are compatible with cholangiocytes (and cholangiocyte senescence) playing a central role in PSC, as previously mentioned: (1) the PSC-microbiota hypothesis ^, and (2) the gut lymphocyte homing hypothesis. The PSC-microbiota hypothesis is based, in part, on the association between PSC and IBD and represents an expansion of what has been termed the “leaky gut” hypothesis. It posits that PSC may develop as a result of: (1) increased enterohepatic circulation of microbial molecules (possibly facilitated by compromised intestinal barrier function), (2) alterations in gut microbial diversity and/or the repertoire of metabolites (e.g., because of intestinal microbial dysbiosis), and/or (3) an aberrant or exaggerated cholangiocyte or other hepatic cell response to microbial molecules. This hypothesis is supported by various findings in vitro, in animal models, , and in vivo. ^, Human studies have evaluated the bacterial diversity of stool in healthy controls, patients with PSC, and patients with UC without liver disease; patients with PSC had a unique microbial composition compared with the other two groups and reduced bacterial diversity, although there was no significant difference among patients with PSC with or without IBD.

The gut lymphocyte homing hypothesis postulates that intestinal T lymphocytes are: (1) activated in gut-associated lymphoid tissue, (2) primed by dendritic cells to express the cell surface receptors integrin α4β7 and CCR9, and (3) recruited to the liver as a result of aberrant hepatic expression of their cognate ligands, namely the addressin protein MAdCAM-1 and the chemotactic protein CCL25, which are usually restricted to the intestine. ^, Although the hepatic expression of these ligands on periportal endothelial cells and subsequent homing of α4β7+, CCR9+ lymphocytes to the liver appears to be relatively specific to PSC, ^, the pathobiologic relevance of this process has not been well-defined but is believed to represent a means of initiating peribiliary inflammation and cholangiocyte injury. The discovery of α4β7 as a possible therapeutic approach led to the human studies of vedolizumab, which will be discussed later in this chapter.

Genetic factors appear to play a role in the development of PSC or in modifying its phenotype and may well interconnect with the aforementioned hypotheses. Several lines of evidence support a role for genetic factors in PSC: first, the risk of PSC is significantly increased within offspring and siblings of patients with PSC (hazard ratio of approximately 11). ^, Second, data from genome-wide association studies suggest that the human leukocyte antigen (HLA) gene family collectively represents the strongest risk locus associated with PSC. Third, various non-HLA susceptibility and modifier genes have been identified, including but not limited to stromelysin-1 (i.e., matrix metalloproteinase 3), intracellular adhesion molecule 1 (ICAM1) , and matrix metalloproteinase (MMP) 1 and 3.

There have been numerous mouse and rat models developed to study various features of PSC. Given the uncertainties regarding the etiopathogenesis of PSC, it is not surprising that no single animal model has recapitulated all of the biochemical, cholangiographic, histologic, and pre-malignant features of PSC. Indeed, although the Mdr2 ^-/- ( ABCB4 gene) knockout mouse has been the most widely studied model and exhibits biochemical, histologic, and cholangiographic features of the human PSC ( Fig. 41.4 ), it is discordant from PSC in several aspects. For instance, disease severity in the Mdr2 ^-/- mouse appears to be worse in female mice, and there is no apparent association with IBD or CCA. ^, Moreover, the mechanism of injury in the model (decreased biliary phospholipids resulting in hydrophobic bile-mediated epithelial injury and leakage of bile into portal tracts) has not been shown to correspond to human PSC. Thus there is no consensus to date regarding the optimal animal model, and this has hindered development and testing of potential therapies for PSC.

Natural history

PSC is a chronic disorder that generally progresses to end-stage liver disease, even among asymptomatic patients. The median LT-free survival from the time of PSC diagnosis is approximately 20 years. ^{, ,} LT-free survival in children with PSC appears to also be within this range. ^, It should be noted that data from large academic centers may suggest shorter median survival as a result of referral bias, whereas early diagnosis of PSC may be associated with longer survival because of lead-time bias. ^{, ,}

Prognostic models for PSC have been developed to predict survival and hepatic decompensations and to identify the ideal timing of LT. ^, A commonly used prognostic model, which does not require liver histologic data, is the revised natural history model for PSC (i.e., the revised Mayo PSC risk score, which utilizes patient age, serum bilirubin, albumin, aspartate aminotransferase, and history of variceal bleeding). Similarly, the Model for End-Stage Liver Disease (MELD) and MELD-Na are used for LT allocation for PSC and non-PSC patients and are a surrogate marker for mortality risk ^, (see Chapter 4 ). Normalization of ALP has been evaluated in several studies as a biomarker of improved survival and decreased risk of major adverse PSC-related events but continues to be a subject of uncertainty and ongoing investigation. The limitations of current prognostic models are several, including, for instance, their inaccuracy in predicting development of CCA and impairment of health-related quality of life; indeed, no consensus exists regarding the optimal model.

Patients with small-duct PSC (5% of all PSC), a term that refers to patients with biochemical and histologic features consistent with PSC but a normal cholangiogram ( Table 41.2 ), appear to have a better long-term prognosis compared with those with non–small-duct PSC. In a multi-institutional and multi-national study, 83 patients with small-duct PSC were matched by age, sex, year of diagnosis, and institution in a 1:2 fashion to patients with large-duct (i.e., classic) PSC. Patients with small-duct PSC did not develop CCA unless their disease progressed to large-duct PSC (which occurred in 23% over a median of 7.4 years) and had significantly longer LT-free survival (13 compared with 10 years).