• Open Access

Ascites is a Negative Prognostic Indicator in Chronic Hepatitis in Dogs


  • This work was performed at the Department of Veterinary Medicine, University of Cambridge, Cambridge, UK. Preliminary data were presented as a clinical research abstract at the ECVIM-CA Congress September 13–15, 2007 in Budapest, Hungary.

Corresponding author: Eleanor Raffan, Queen's Veterinary School Hospital, Department of Veterinary Medicine, University of Cambridge, Madingley Road, Cambridge CB3 0ES, UK; e-mail: er311@cam.ac.uk.


Background: Chronic hepatitis (CH) in dogs is common but little is known about factors associated with survival. Ascites is a well-recognized negative prognostic indicator in humans.

Hypothesis: Ascites is a negative prognostic indicator in CH in dogs.

Animals: Thirty-four dogs with histologically confirmed CH presented to 1 institution between 1996 and 2005.

Methods: Retrospective observational study. CH was diagnosed by histopathology of liver tissue according to the WSAVA criteria. Ascites was diagnosed by abdominal ultrasound. The association of ascites with survival from diagnosis or onset of owner-reported clinical signs until death from any cause or from liver disease was analyzed. Ascitic and nonascitic groups were further analyzed for differences in treatment and sex.

Results: Fourteen of 34 dogs had ascites. Survival from diagnosis to death from liver disease was 0.4 months (95% confidence interval [CI], 0.2–0.6) for ascitic dogs and 24.3 months (CI 11.4–37.1) for nonascitic dogs (P < .001), and from onset of signs to death from liver disease was 2.0 months (CI 0.0–5.6) for ascitic dogs and 33.0 months (CI 8.6–57.4) for nonascitic dogs (P= .0020). Diet and spironolactone use differed between groups.

Conclusions and Clinical Importance: Ascites is a significant negative prognostic indicator in dogs with CH. Veterinarians and owners can use this information to aid clinical decision making in affected dogs.

Chronic hepatitis (CH) is a well-recognized problem in canine practice and affects dogs of many breeds.1–5 Factors affecting prognosis have been the subject of a number of studies2,6–8 which have examined various clinicopathologic, histologic, and clinical variables. None, however, have identified a single factor that clearly affects survival. Ascites is a common clinical sign in CH2,3 and can occur as a result of portal hypertension, neurohormonal activation,9,10 and hypoalbuminemia.11

Ascites in humans has long been recognized as an indicator of poor survival.12–14 In this study, we hypothesized that ascites is a negative prognostic indicator in dogs with CH. This retrospective cohort study of survival in dogs with CH compared those with and without ascites and examined for possible confounding factors that were different between the groups.


Case Selection

The records of the diagnostic histopathology service of the Queen's Veterinary School Hospital, Department of Veterinary Medicine, University of Cambridge were reviewed for cases of canine liver disease admitted between January 1, 1996 and December 31, 2005. All dogs with histopathologic diagnoses suggestive of CH on original biopsy reports were selected and the original sections were reviewed by 1 author (TJS) who was blinded to the clinical details. All cases included were confirmed to have CH according to the WSAVA histologic criteria.15 Biopsies were performed either percutaneously under ultrasound guidance, by wedge biopsy during exploratory laparotomy or at postmortem.

Cases were included if there were documented increases in alanine aminotransferase, clinical signs compatible with liver disease at the time of biopsy, clinical records were available and the clinical outcome known. Dogs were excluded from the study if an underlying cause of liver disease (including infection, toxicity, or copper storage abnormalities) were identified in clinical records, if follow-up data were not available, or if there was significant nonhepatic concurrent disease at the time of biopsy.

Signalment, drug treatment, date of onset of clinical signs, and outcome were recorded. The date of onset of clinical signs was identified by reviewing case records of clinical history taken at admission. Outcome, including date and suspected cause of death, was identified from clinical records or telephone interview with owners or referring veterinarians. The censor date for animals recorded as still alive was March 31, 2007. Ascites was defined as the presence of free abdominal fluid identified on abdominal ultrasound examination or at laparotomy.16 Ascites was recorded as being present or absent at the time of biopsy.

Statistical Analysis

All statistical analyses were performed with a statistical software programme.a Descriptive statistics were used to describe the data distribution. Kaplan-Meier survival analysis17 was performed to analyze the association of ascites status with measured outcomes. For each case, start point and measured outcome were defined in 2 ways. Start point was defined as either onset of clinical signs or time of biopsy diagnosis. The outcome was defined as either conservative survival time (time to death of whatever cause) or liver survival time (time to death from liver disease). This approach yielded 4 analyses for each variable. Survival was considered different between groups when the log-rank statistic18 yielded P < .05. Median survival time (with 95% confidence limits) was recorded for each analysis.

Fisher's exact test19 was used to examine if there was an association between ascites status and either sex or the use of individual treatments (prednisolone, antibiotics, lactulose, ursodeoxycholic acid, S-adenosylmethionine, colchicine, spironolactone, gastric protective drugs, and diet). An association was deemed significant if P < .05.


Forty-two dogs were identified with histopathologically confirmed CH. Seventeen of 42 samples were wedge biopsy sections (2 from necropsy specimens) and the remaining 25 were obtained by needle biopsy. One dog (a Bedlington Terrier with copper storage disease) was found to have an underlying cause for the hepatitis, 2 had significant concurrent diseases, and 5 were lost to follow-up. As a result, 34 dogs met the inclusion criteria for the study, of which 14 had ascites at biopsy diagnosis. Fluid was confirmed to be a transudate or modified transudate by laboratory analysis in 11 of the 14 ascitic dogs. Small volumes of abdominal fluid precluded sampling in the remaining 3 dogs. Nineteen breeds were represented with the largest groups being 7 Labrador Retrievers, 3 Springer Spaniels, and 3 Great Danes. Ages ranged from 1 to 13 years (mean 5 years). There were 18 female dogs and 16 male. Five dogs were euthanized owing to diseases unrelated to CH (3 nonliver tumors, 1 degenerative joint disease, and 1 megaesophagus). Of the remaining 29 dogs, 22 were euthanized, 4 died for reasons related to CH, and 3 were alive at the end point of the study. Postmortem examinations were carried out in 2 dogs.

Survival times were significantly longer for nonascitic dogs than for ascitic dogs with each of the different start and end points (Table 1). Kaplan-Meier survival plots generated with the liver survival end point are presented for data from onset of clinical signs to death (Fig 1) and biopsy diagnosis to death (Fig 2).

Table 1.   Median survival times (95% confidence limits) in months for ascitic and nonascitic dogs.
Survival Time
from (months)
Statistic P
Onset of clinical signs
 Conservative2 (0.0–5.7)20 (4.7–35.3).014
 Liver2 (0–5.7)33 (8.6–57.4).0020
Biopsy diagnosis
 Conservative0.4 (0.0–0.6)14.8 (2.5–27.0).0010
 Liver0.4 (0.2–0.6)24.3 (11.4–37.1)<.001
Figure 1.

 Kaplan-Meier survival plot for nonascitic (continuous line) and ascitic (broken line) dogs from biopsy diagnosis to death from liver disease. Censored cases (died of nonliver disease causes) are identified with a cross.

Figure 2.

 Kaplan-Meier survival plot for nonascitic (continuous line) and ascitic (broken line) dogs from onset of clinical signs to death from liver disease. Censored cases (died of nonliver disease causes) are identified with a cross.

Fisher's exact test was applied to see if there was a difference between ascitic and nonascitic groups. There was no difference between groups with respect to sex. Incomplete information on treatment was available for 3 dogs in each of the ascitic and nonascitic groups. There was no difference between groups with respect to frequency of treatment with prednisolone, antibiotics, lactulose, ursodeoxycholic acid, S-adenosylmethionine, colchicine, or gastric protective drugs. Dietary therapy was more likely to be used in the nonascitic group (13/18 dogs in the nonascitic group as opposed to 3/13 in ascitic group, P= .01). Spironolactone was more commonly used in the ascitic group (3/18 dogs in the nonascitic group as opposed to 7/13 in ascitic group, P= .05). All dogs in the nonascitic group that received spironolactone did so when they developed ascites after biopsy diagnosis of CH.


This study reveals that ascites is a strong negative prognostic indicator in dogs with CH. Ascites formation in CH has been studied extensively in humans and with naturally occurring and experimental canine models of liver disease. The inciting event in the development of ascites is activation of hepatic stellate cells (Ito cells) with subsequent perisinusoidal collagen accumulation and fibrosis that results in sinusoidal hypertension followed by portal hypertension.20 Neuronal and hormonal activation are also involved in the development of ascites.9,10 Hypoalbuminemia, causing reduced plasma oncotic pressure, may contribute to the ascites of liver disease but is rarely sufficient to cause ascites on its own.11

Ascites may be associated with poor survival in dogs for a number of reasons. It may develop when there is more severe liver pathology, or it may be that the associated neuroendocrine abnormalities increase morbidity, or the distended abdomen resulting from ascites may prompt owners to request euthanasia earlier. This latter reason is unlikely to account for the entire effect given that ascites has long been recognized as a negative prognostic indicator in human liver disease patients.12–14,21 In human patients, development of ascites is regarded as an important “landmark” in the progression of cirrhosis and triggers requests for urgent orthotoptic liver transplant in eligible patients.21,22 A recent veterinary study of dogs with presinusoidal portal hypertension of variable etiology also identified a poor prognosis. Of the 17 dogs in the study, 7 had CH.23

The data were examined for confounding factors that could have altered the conclusions. There was no difference between the ascitic and nonascitic groups with respect to sex, or any treatment except diet and spironolactone. That spironolactone was more commonly used in ascitic dogs is no surprise given that it is the diuretic of choice for treatment of ascites associated with liver disease in canines.24 The fact that dietary management was more commonly used in nonascitic dogs likely reflects both reduced morbidity in these patients (hence better acceptance of a new diet) and the longer survival time in which to implement the dietary change.

Other studies have associated various clinical,7 biochemical,2,6–8 coagulation,7,8 and histopathologic2,8 parameters with differences in prognosis in canine CH. However, none have shown a marked clinical distinction in survival comparable with that seen with ascites in this study.

Every case in the study had histopathologic changes consistent with CH as defined by the WSAVA Liver Standardization Group.15 None had an inciting cause such as copper storage disease, infection, or toxicity identified. CH is often diagnosed late in the disease course, however, and there is considerable debate as to the underlying etiopathogenesis.2,3,5,15,20,25 As a result, our cases may represent a heterogeneous population with different inciting causes. While future understanding may help subclassify dogs with CH, the group was representative of cases seen in clinical practice and hence the findings are relevant to veterinarians working today.

This study has identified ascites as a negative prognostic indicator in canine CH and reported survival times that will be helpful to both clinicians and pet owners when managing dogs with CH. While prognostic indicators can be helpful, it is important to note that some dogs with ascites had protracted survival so the presence of ascites in a dog who is otherwise coping well with liver disease should not prompt euthanasia. The retrospective nature of this study meant we were not able to examine if ascites refractory to treatment carried a worse prognosis but this would be an interesting route of investigation in future.


aSPSS 16.0, SPSS Inc, Chicago, IL


The authors thank Dr Fred Heath for help with the statistical analysis. PJW's senior lectureship is funded by the Bunning Endowment and the Iams Company and ER's residency is funded by the Alice Noakes Memorial Trust.