High frequency of cholestasis in generalized pustular psoriasis: Evidence for neutrophilic involvement of the biliary tract


  • Manuelle Viguier,

    Corresponding author
    1. Department of Dermatology 1, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    • Service de Dermatologie I, Hôpital Saint-Louis, 1 Avenue Claude-Vellefaux, 75475 Paris Cedex 10, France
    Search for more papers by this author
    • fax: +33 1-42-49-46-20

  • Matthieu Allez,

    1. Department of Gastroenterology, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Anne-Marie Zagdanski,

    1. Department of Radiology, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Philippe Bertheau,

    1. Department of Pathology, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Eric de Kerviler,

    1. Department of Radiology, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Michel Rybojad,

    1. Department of Dermatology 2, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Patrice Morel,

    1. Department of Dermatology 2, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Louis Dubertret,

    1. Department of Dermatology 1, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Marc Lémann,

    1. Department of Gastroenterology, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author
  • Hervé Bachelez

    1. Department of Dermatology 1, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris and Université Paris 7, Paris, France
    Search for more papers by this author


Generalized pustular psoriasis is a rare form of psoriasis that is sometimes associated with extracutaneous manifestations. Evidence for biliary involvement has been suggested in isolated cases. We investigated the prevalence and nature of liver abnormalities occurring in this disease. Twenty-two patients consecutively admitted for generalized pustular psoriasis who underwent liver biological tests at the time of the attack and during the following weeks were included. Twenty patients (90%) had at least one abnormal biological liver parameter. Eleven patients (50%) had pronounced abnormalities: jaundice (4/22), gammaglutamyl transferase higher than 5 times the normal value (10/22), alkaline phosphatase higher than twice the normal value (7/22), and/or aminotransferases higher than 3 times the normal value (7/22). These abnormalities returned to normal range at the time of remission of pustular psoriasis, suggesting that severe liver abnormalities could be associated with severe cutaneous disease. Neutrophilic cholangitis was observed on liver biopsy. Persistent magnetic resonance cholangiopancreatography features similar to those observed in sclerosing cholangitis were found in 3 of the 4 patients studied. No causal factor other than pustular psoriasis could be identified. In conclusion, our results demonstrate the high frequency of liver abnormalities in patients with generalized pustular psoriasis. Biliary involvement related to neutrophilic cholangitis should be added to the spectrum of extracutaneous manifestations of this disease, and physicians should be aware of such a complication. (HEPATOLOGY 2004;40:452–458.)

Generalized pustular psoriasis (GPP) of the von Zumbusch variety is a generalized life-threatening form of psoriasis characterized by disseminated pustular skin lesions following an acute course and often associated with high fever and hyperleukocytosis. The attacks may be induced in some cases by triggering factors such as infections, drugs, sunlight, or pregnancy.1, 2 Since the first description of the disease in 1910, there have been reports of patients with GPP who present with extracutaneous manifestations, which consist mostly of mucous membranes lesions and arthritis.3, 4 Evidence of liver involvement of unknown physiopathology has been suggested in isolated case reports.5–8 Recently, we reported a case of neutrophilic cholangitis associated with GPP in a 42-year-old woman presenting with cholestasis following an acute evolution paralleling the course of skin disease.9 Recognizing the hypothesis that neutrophilic cholangitis could represent an extracutaneous manifestation of GPP, we investigated the prevalence and the nature of liver abnormalities occurring in patients with GPP.


GPP, generalized pustular psoriasis; CRP, C-reactive protein; GGT, γ-glutamyltransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ERCP, endoscopic retrograde cholangiopancreatography; MRCP, magnetic resonance cholangiopancreatography; PSC, primary sclerosing cholangitis.

Patients and Methods

We conducted a study on all patients consecutively admitted for a GPP attack in one of the two departments of dermatology of the Saint-Louis Hospital, Paris, between 1995 and 2001. To be included, patients had to fulfill the following criteria: (1) at least one attack of GPP, defined as sheets of erythema and pustulation spreading in waves, progressively involving the whole body surface within a few days, associated with general symptoms including fever higher than 38°C, and/or leukocytosis higher than 10,000/mm3; or (2) liver biological tests available at the time of GPP attack and during the following weeks. The clinical criteria were validated by two dermatologists, and the diagnosis of GPP had to be ascertained on the basis of the histological analysis of a skin biopsy showing changes consistent with pustular psoriasis (i.e., acanthosis, parakeratosis, and spongiosis with subcorneal and intraepidermal collections of neutrophils).

Extracutaneous clinical manifestations during the attack were noted, including fever, mucous membrane lesions, abdominal pain, jaundice, arthritis, and pneumonitis. Biological investigated parameters were the value of peripheral blood absolute leukocyte count, C-reactive protein (CRP), bilirubin, γ-glutamyltransferase (GGT), alkaline phosphatase (ALP), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) serum levels during the acute episode. These parameters were considered abnormal if the highest value was above the normal upper limit.

A subset of patients with abnormal liver tests was extensively investigated: search for drugs intake, serological detection of hepatitis B virus and hepatitis C virus infection, abdominal ultrasound examination, liver histology, endoscopic retrograde cholangiopancreatography (ERCP) or magnetic resonance cholangiopancreatography (MRCP). Because histological analysis of liver samples gave confirmation of a specific involvement in two initial cases, liver biopsy was not performed in further enrolled patients because of potential complications associated with this invasive technique. Moreover, ERCP was substituted in the last studied patients by MRCP, which is a noninvasive technique previously shown to be both as sensitive and specific as ERCP for the detection of extrahepatic bile duct lesions in diseases such as primary sclerosing cholangitis (PSC).10–15 MRCP was performed using a 1.5-T unit (Siemens Magnetom Symphony, Erlangen, Germany). Heavily T2-weighted sequences (single-shot turbo spin echo) were used with multiple coronal oblique thick slabs (3- to 5-cm section). Contiguous thin (7-mm) slices were also obtained in the axial plane using the same parameters. Results from MRCP were independently analyzed by two radiologists.

The evolution of clinical and biological liver abnormalities, in relation to skin and general symptoms of GPP, were assessed, as well as the outcome of liver involvement after cutaneous healing and during further relapses of GPP.

Statistical analysis was performed using Student's t test or linear regression analysis with Statview software (SAS Institute, Cary, NC).


Characteristics of Patients and Psoriasis History.

From an initial population of 26 patients admitted for a GPP attack, a total of 22 patients were enrolled. The 4 remaining patients were excluded because results from liver tests were lacking. There were 16 women and 6 men (sex ratio F/M: 2.7). The median age, at the time of the studied GPP attack, was 45 years (range, 21–98). While the studied GPP attack was the first one in the psoriasis history of most patients, 7 of them had experienced at least one attack prior to inclusion. Pustular psoriasis attacks occurred as an evolutive complication of the usual clinical variety of psoriasis in 17 patients (77%): psoriasis vulgaris in 12 of them, acral pustular psoriasis (pustular palmoplantar psoriasis or acrodermatitis continua) in 3 cases, localized pustular psoriasis (1 patient), and erythrodermic psoriasis (1 patient). Four other patients suffered from GPP of the von Zumbusch type since childhood, while the GPP attack occurring at the time of inclusion was the initial manifestation of psoriasis history in 1 patient.

Extracutaneous Manifestations Observed During Attacks.

Twenty-one patients (95.4%) had a severely altered general condition, with malaise, chills, and a high-grade fever (median, 39°C; range, 37°C–40°C). Extracutaneous manifestations were observed in 13 patients (59%) including epigastric pain in 4, jaundice in 4, polyarthritis in 3, interstitial pneumonitis in 2, mucous membrane lesions involving the tongue or the gums in 2, otitis media in 1, aseptic pyuria in 1, and aseptic leucorrhea in 1. All these extracutaneous manifestations were transient and remitted in parallel with cutaneous involvement.

Laboratory Investigations.

Nineteen patients (86%) had a neutrophil absolute count exceeding 10,000/mm3 (median, 16,650/mm3; range, 7,600–33,900). Considering the highest biological value observed during the attack of GPP compared with the upper limit of normal value at our laboratory (N), 20 of the 22 patients (90%) had at least one abnormal biological liver parameter. Seventeen patients (77%) had an abnormal GGT level (median, 3.3 N; range, 1–14.6 N). An abnormal ALP value was observed in 13 patients (59%; median, 1.2 N; range, 1–5.6 N), while abnormal AST and ALT values were found in 12 patients (median AST count, 1.2 N, range, 1–14 N; median ALT count, 1.5 N, range, 1–22.3 N). The bilirubin value was higher than 16 μmol/L, which is the normal limit at our laboratory, in 8 patients (36%; median, 1; range, 1–14 N) (Table 1). Pronounced liver abnormalities were observed in 11 out of 22 patients (50%): bilirubin superior to 50 μmol/L (4 patients), GGT levels higher than 5 times the normal value (10 patients), ALP values higher or equal to twice the normal value (7 patients), and/or aminotransferase (AST and/or ALT) values higher or equal to 3 times that of normal (7 patients) (see Table 1). No statistical difference was evidenced regarding age, gender, or presence of other extracutaneous manifestations between these patients and the remaining ones (Student's t test, data not shown).

Table 1. Biological Variables, Histological and Morphological Findings in Studied Patients
PatientNeutrophil Count (/mm3)Bilirubin Count (μmol/L)GGTALPASTALTLiver BiopsyERCPMRCP
  1. NOTE: The highest value observed during the episode has been selected and compared with the normal upper limit (N). —, not done.

  2. Abbreviations: CBD, common bile duct; PBD, peripheral bile duct.

1886.4 N3 N3.4 N2 NNeutrophilic cholangitisNormal
218,10067.18.9 N5.6 N2.6 N3.35 NNeutrophilic cholangitisUnique stricture of CBD
319,10045.86.3 N2.1 N2.3 N1.9 NNormal
421,0008 N3 N1.5 N
519,900221.57.8 N1.8 N14 N22.3 N
617,000548 NN2.8 N3.6 N
733,900<162 NNNN
829,000<163.4 N1.2 N1.2 NN
97,60022.8NN2.1 NN
1015,100<161.1 NNN1.6 N
1114,600<162.2 N2.3 N2.2 N3.3 N
1221,500<161.6 NNN1.5 N
139,700<16N1.1 NNN
1410,400<163.25 NNNN
1511,500<165.2 N2 N1.6 N1.6 N
1619,100<162.5 NNNN
1716,20025.911.5 N5.5 N3 N4 N
1916,300<166.8 N1.4 N1.3 NNStrictures of PBD
2016,70021.614.6 N1.5 N1.2 N3 NStrictures of CBD and PBD
2222,400<16N1.2 NNN

We further investigated whether or not a relationship could be established between the severity of liver/biliary alterations and the severity of cutaneous disease. The severity of GPP could not be directly assessed because the psoriasis area severity index is not useful in this variety of psoriasis.16 We studied in all patients the correlation between different parameters reflecting the severity and/or the duration of the GPP attack (highest temperature recorded, maximum values of neutrophil count and of CRP, duration of hospitalization) and the highest value of the different biochemical liver tests (bilirubin, GGT, ALP, AST, and ALT values). A significant correlation was found between CRP and ALP values (r = 0.616, P = .01) and between CRP and bilirubin levels (r = 0.551, P = .05). In contrast, there was no significant correlation between biochemical liver tests and fever, duration of hospitalization, or neutrophil count.

The maximal peripheral blood leukocyte absolute count was observed 6 days after the onset of the GPP attack (range, 1–15), while the peak value for liver abnormalities occurred at day 6 for bilirubin (range, 2–46), day 11 for ALP (range, 7–23), day 15 for GGT (range, 6–40), day 16.5 for ALT (range, 5–40), and day 16 for AST (range, 11–44). For all of these values, the median one is given.

Investigations in Patients With Liver Alterations.

All but 2 patients (cases 16 and 21) required systemic treatment for GPP attack consisting in acitretin (n = 13), cyclosporine (n = 5), methotrexate (n = 3), or isotretinoin (n = 1) taken alone or in combination. Drug-induced toxicity was excluded in 17 out of 20 patients (85%) because liver abnormalities were observed prior to the intake of these drugs and returned to normal ranges despite continuation of drug therapy.

Serologies for hepatitis B virus and hepatitis C virus were performed in 15 out of the 20 patients having at least one abnormal biological liver parameter; all but one case yielded negative results. This latter case (patient 5) was positive for hepatitis B surface antigen but negative for hepatitis B virus deoxyribonucleic acid, excluding hepatitis B virus infection as a causal factor for liver abnormalities. Abdominal ultrasonography was performed in 14 patients and was normal in 8 (57%). Homogenous hepatomegaly with features of steatosis was observed in 3 cases, whereas uncomplicated vesicular lithiasis was shown in 1 case. A slight dilatation of the biliary tract and a thickened biliary tract were observed in patients 3 and 19, respectively.

ERCP was performed in one case and was normal (patient 1). Four patients underwent MRCP during one of the acute episodes. MRCP correctly showed the common bile duct in all cases and the intrahepatic bile ducts in all but one case. Bile duct strictures—defined as narrowing of the bile duct with dilatation of the proximal segment—were found in 3 of the 4 patients. One of them had a unique stricture of the common bile duct (Fig. 1A). Another one had 2 strictures involving the common bile duct and multiple peripheral bile duct strictures (Fig. 1B and 1C). The remaining patient had 3 peripheral bile duct strictures (Fig. 1D).

Figure 1.

Bile duct strictures in 3 patients (MRCP). (A) Moderate narrowing of the common hepatic bile duct (arrow) in patient 2. (B, C) In patient 20, MRCP shows two strictures of the common hepatic bile duct (arrows) and succession of strictures (arrowheads) and dilations of intrahepatic bile ducts (B, left view; C, right view). (D) A stricture of the left central hepatic bile duct (arrow) with proximal dilatation and irregularities of the left intrahepatic bile ducts (arrowhead) are observed in patient 19.

Two patients (patients 1 and 2) with pronounced abnormalities on liver tests underwent a liver biopsy at the time of GPP attack. Histological examination showed no architectural change. In the two cases, a polymorphonuclear leukocyte infiltrate was observed in the portal connective tissue as well as within the epithelial layer of interlobular bile ducts, without epithelial cell degeneration (Fig. 2). No polymorphonuclear leukocyte was seen in the bile duct lumen or in the hepatic lobules. There was no cholestasis, no portal edema, and no periductal fibrosis. A mild bile ductular proliferation was observed in patient 1. Associated liver abnormalities consisted of mild lymphocytic portal infiltration, mostly composed of CD8+ T cells, without lymphocytic bile duct exocytosis. There were few liver cell necroses in periportal areas and a mild steatosis. Recently, another patient was directly referred to our gastroenterology department from an external dermatology department for jaundice, acute abnormal liver tests, and normal abdominal ultrasound examination during a GPP attack. We did not include this patient in the present series, because the recruitment method was different. Liver biopsy revealed alterations very similar to those described above. A mild bile ductular proliferation was also observed in this patient (data not shown).

Figure 2.

Histological views of portal tracts showing neutrophilic cholangitis in 2 patients. Polymorphonuclear leukocytes are observed in the portal spaces, as well as within the epithelial layer of interlobular bile ducts (arrows), but not in the bile duct lumen or within the hepatic lobules (hematoxylin-eosin; original magnification ×400).

Outcome of Liver Alterations.

Liver tests returned to normal values within the month following the peak value after recovery of the cutaneous lesions in all but 1 patient. In all 6 patients who presented with further GPP attacks during follow-up, a relapse of liver alterations with a similar course was observed. The 3 patients who had abnormalities evidenced by MRCP underwent a follow-up MRCP examination 6, 9, and 27 months after the initial one, respectively. Abnormal features were persistent in all cases.


GPP of the von Zumbusch variety is a severe form of psoriasis that may be associated with life-threatening complications such as sepsis or metabolical, hemodynamic, and thermoregulatory disturbances related to severe alterations of the epidermal barrier. Although cutaneous involvement is always the predominant feature, extracutaneous localizations have also been described, and liver involvement has only been reported in isolated cases prior to this study.5–8, 17–23 However, the incidence of liver damage has not been previously investigated in published series of patients with GPP.1, 2, 24–27 The results presented herein provide evidence for a high frequency of liver involvement in patients with GPP. Indeed, we have shown that clinical symptoms (epigastric pain and/or jaundice) were present in 32% of cases, while the prevalence of at least 1 abnormal biological liver test was 90%. Pronounced abnormalities on liver tests were observed in 50% of patients. Liver test abnormalities usually occurred within the 2 weeks following the onset of cutaneous symptoms and 1 week after the peak of leukocytosis. A significant correlation was found between CRP and ALP values at one hand and between CRP and bilirubin levels on another hand. This suggests that the severity of liver alterations could be associated with a more severe cutaneous flare. Nevertheless, this needs to be confirmed in further prospective studies designed to assess both inflammatory biological markers and different clinical parameters that were not investigated in the present study (e.g., the duration of pustulation and erythroderma).

Abnormal liver tests usually resolved when the skin healed and relapsed during further GPP attacks in all cases. The evidence that the course of liver alterations paralleled the evolution of both skin lesions and neutrophilic hyperleukocytosis strongly supports a specific involvement of the biliary tract. No other cause of liver and/or biliary disorder such as viral infection, toxic shock syndrome, cardiac failure, or obstruction or infection of the biliary tree were found in our patients. No potentially hepatotoxic drug, such as acitretin or methotrexate, was taken by our patients before occurrence of liver abnormalities in 85% of them (data not shown). Biliary involvement can be considered an extracutaneous manifestation of GPP in view of histological alterations observed in liver biopsies from 3 patients showing a neutrophilic infiltration of the portal tract and of the epithelium of the bile ducts, which is consistent with the diagnosis of neutrophilic cholangitis. MRCP revealed intra- or extrahepatic bile duct strictures in 3 out of 4 patients analyzed. These latter features were still observed several months after the initial investigation in these 3 patients. Biliary involvement related to GPP was not limited to small ducts, but also affected large biliary ducts with strictures, sharing features similar to those observed in PSC. Interestingly, a case of GPP has been previously mentioned in a series of 56 children with PSC.28 However, no histological changes characteristic of PSC (e.g., concentric layers of connective tissue surrounding bile ducts29) or changes associated with large duct obstruction (e.g., cholestasis, portal edema, or presence of polymorphonuclear leukocytes in the bile duct lumen29) were observed in the 3 liver biopsies we examined, one of them having being performed in a patient who had a unique stricture of the common bile duct (case 2). Moreover, patients with stenosis on MCRP did not develop progressive cholestasis, as liver tests remained normal between GPP attacks with a follow-up now exceeding 4 years for patient 2. Nevertheless, specific pathological signs of PSC are seen in only a minority of percutaneous liver biopsies, and it is more common to find nonspecific fibrosis and inflammation in the portal tracts, especially at the early onset of disease.29 On the other hand, there are reports of patients with PSC diagnosed by cholangiography or MRCP who had normal liver tests.29, 30 Altogether, the presence of PSC in GPP patients remains uncertain and follow-up studies will be warranted to clarify further the prognosis of biliary duct stenosis in patients with GPP.

Case reports of liver abnormalities occurring during GPP attacks have been previously published, consisting mostly of liver test abnormalities with or without jaundice.5, 6, 8, 17–23 The hypothesis that liver abnormalities in patients with GPP could be related to the inflammatory pathogenic process characterized by a predominant neutrophilic activation has already been raised by several authors.5, 6, 8 In the reported cases, ultrasonography, tomodensitometry explorations,5, 8, 22 ERCP,8 or MRCP22 showed no significant change. However, a specific involvement was supported by histological alterations found by liver biopsy which was performed in four cases, showing evidence of cholestasis with portal fibrosis and proliferation of ductules,8 periportal infiltrates of polymorphonuclear leukocytes, and lymphocytes with focal necrosis of liver cells.5, 6 These alterations differed from histopathological findings of our series regarding the lack of duct epithelial cell damage, but these differences may reflect different stages during the course of liver involvement.

Neutrophilic infiltration, as found in the portal tract of our patients, is a predominant feature of cutaneous and extracutaneous lesions of pustular psoriasis, notably those involving the mucous membranes3 and the synovial membranes in patients with polyarthritis.4 Other extracutaneous manifestations of GPP include ocular involvement resulting in conjunctivitis, uveitis, or corneal ulceration,31, 32 as well as bone lytic lesions due to multifocal sterile osteomyelitis.33 Pneumonitis with occasionally acute respiratory distress syndrome has also been reported, but neutrophilic involvement of the bronchial tract has not been documented in published cases.19, 34–37 Interestingly, one case of neutrophilic cholangitis associated with Sweet's syndrome—a variety of neutrophilic dermatosis—has also been reported.38

So far, it remains unclear whether neutrophilic infiltration of the biliary tract is an initial event or more likely is secondary to the influx of activated T cells and macrophages, as suggested by Shelley5 and Craig,6 in a similar way as what is observed for skin lesions.39 In future studies, immunohistochemical analysis of T cell subsets and investigations of the local secretion of neutrophil-recruiting chemokines (e.g., interleukin 8 and Gro-α) will be warranted to deal with this issue.

In conclusion, the results of the present study strongly support that neutrophilic cholangitis should be added to the spectrum of extracutaneous manifestations of GPP. Such a complication should be known to avoid both useless invasive liver investigations and withdrawal of drugs with potentially deleterious consequences on the course of the disease. Further prospective studies are warranted to understand more accurately the mechanism of biliary involvement in GPP.


The authors are indebted to Dr. Ewa Guigné (Department of Dermatology, Hôpital Henri Mondor, Créteil, France) for referring a patient.