Concurrent hepatopathy in dogs with gallbladder mucocele: Prevalence, predictors, and impact on long-term outcome

1 INTRODUCTION

Gallbladder mucocele (GBM) is an increasingly recognized and important cause of biliary tract disease in dogs.^1-4 The disease is characterized grossly by accumulation of inspissated, immobile mucus and bile in the gallbladder (GB), and histologically by hyperplastic GB mucosa that forms cystic spaces filled with mucus and papillary fronds that project into the GB lumen.⁵ The specific etiopathogenesis remains unclear. Cholecystectomy is currently considered the treatment of choice for dogs with GBM.⁴

A variety of complications have been associated with GBM including but not limited to GB necrosis and rupture, extra-hepatic biliary duct obstruction (EHBDO), bacterial infection, cholecystitis, coagulation disturbances, and systemic inflammatory response syndrome.^6-13 The impact of a number of these factors on short-and long-term outcomes in dogs with GBM undergoing cholecystectomy have been investigated in multiple studies.^{8, 10, 14, 15} In contrast, concurrent hepatic abnormalities in dogs with GBM are less well described. One retrospective study describes 30 cases of GBM, 22 of which had liver biopsy performed. Hepatic abnormalities were common and included portal inflammation, portal fibrosis, and bile duct hyperplasia.¹² Another retrospective case series described concurrent liver abnormalities in 94.4% of liver samples in dogs with GBM.¹¹ Additionally, no study evaluating short or long-term prognosis in dogs with GBM has evaluated the relationship between hepatic histologic abnormalities and survival.^{8, 15, 16} Outcomes in dogs surviving the post-operative period after cholecystectomy for the treatment of GBM are generally considered good, however death attributed to ongoing hepatobiliary disease is reported.¹¹

The plasma neutrophil-to-lymphocyte ratio has been evaluated as a biomarker for a variety of inflammatory diseases in dogs including meningoencephalitis of unknown etiology, chronic enteropathy, systemic inflammatory response syndrome, pneumonia, acute pancreatitis, and septic peritonitis.^17-23 If inflammatory lesions of the liver were associated with worse outcomes in dogs with GBM, it would be helpful to have a noninvasive biomarker to predict them.

The objectives of this study were to (a) report the prevalence of hepatic histologic abnormalities in a larger cohort of dogs with GBM, (b) evaluate for an association between hepatic abnormalities and long-term outcome in dogs with GBM, and (c) investigate whether the neutrophil-to-lymphocyte ratio (NLR) differs in dogs with GBM with and without hepatic inflammatory lesions. We hypothesized that hepatic histologic abnormalities would be common in dogs with GBM, and inflammatory lesions and fibrosis would be associated with shorter survival. We expected that the NLR would be increased in dogs with GBM and hepatic inflammatory lesions when compared to dogs with GBM and no hepatic inflammatory disease.

2 MATERIALS AND METHODS

A retrospective medical records review was conducted at Michigan State University Veterinary Medical Center and Ozark Veterinary Specialty Care using the search terms “gallbladder mucocele” and “cholecystectomy” from January 1, 2016 to December 31, 2020. Included dogs had a cholecystectomy performed for histologically confirmed GBM with archived gallbladder and liver tissue available for evaluation. Dogs with incomplete medical records or with significant concurrent pathology (eg, neoplasia) were excluded. Recorded data included age, breed, sex, weight, clinicopathologic data, presence or absence of GB rupture, and outcome. The diagnosis of GB rupture was confirmed with review of surgical reports. When complete outcome data were not available in the medical record, referring veterinarians were contacted by telephone to obtain date and reason for euthanasia or death, if known. Outcome data were recorded as alive (1) or dead (0) at 1, 3, and 12 months post-cholecystectomy. To be included, outcome data must have been available for at least the first timepoint (1 month) post-cholecystectomy.

Once cases were selected for the study, histopathologic evaluation of the previously obtained GB and hepatic tissue was performed by a pathologist and pathologist-in-training who were blinded to the case data and outcome (RCM and JAK). First, the diagnosis of GBM was confirmed based on the presence of both cystic mucinous hyperplasia (CMH) of the GB wall and abundant mucus in the GB lumen. Dogs with CMH but lacking accumulation of mucus characteristic of GBM were excluded from the study. Hepatic tissue was then evaluated for the presence or absence of 8 categorical variables: biliary hyperplasia, centrilobular hepatitis, cholestasis, cholangitis, cholangiohepatitis, necrosis, portal inflammation, and vacuolar hepatopathy (Table 1), using WSAVA criteria²⁴ when indicated. Where applicable, the adapted METAVIR histopathology scoring system^{25, 26} (Table 2) was applied in evaluation of fibrosis (eg, Figure 1A-E) and necroinflammatory activity, and individual scores were recorded. Specifically, for lobular necrosis: LN0 = less than 1 necroinflammatory focus per lobule; LN1 = at least 1 necroinflammatory focus per lobule; LN2 = several necroinflammatory foci per lobule or confluent or bridging necrosis.²⁵ For interface hepatitis (previously termed piecemeal necrosis): PMN0 = absent; PMN1 = focal alteration of the periportal plate in some portal tracts; PMN2 = diffuse alteration of the periportal plate in some portal tracts or focal lesions around all portal tracts; PMN3 = diffuse alteration of the periportal plate in all portal tracts.²⁵ Masson's trichrome staining was performed to optimize scoring of fibrosis, and rhodanine staining was performed for qualitative grading of copper according to previously published protocols.²⁷ Finally, the plasma NLR was calculated by dividing the segmented neutrophil count (#/μL) from the lymphocyte total (#/μL).

3 RESULTS

The original search of medical records from January 2016 to December 2020 at Michigan State University Veterinary Medical Center and Ozark Veterinary Specialty Care using the search terms “gallbladder mucocele” and “cholecystectomy” yielded 101 cases. Liver tissue was not available in 1 case. Thirty dogs had CMH but were lacking accumulation of mucus characteristic of GBM. An additional 16 dogs had significant concurrent disease excluding them from the study. Following these exclusions, 54 dogs remained in the study. No outcome data could be obtained for 2 dogs; thus, 52 dogs were included in the study (Figure 2).

Fifteen dogs were castrated males, 2 were intact males, 34 were spayed females, and the remaining dog was an intact female. The median age was 11 (range, 5-16) years. Breeds included mixed breed (13), Shetland sheepdog (4), beagle (3), Labrador retriever (3), West Highland white terrier (3), Chihuahua (2), Jack Russell terrier (2), miniature dachshund (2), shih tzu (2), Yorkshire terrier (2), and 1 each of the following Australian heeler, Australian shepherd, bichon frisé, Brussels griffon, Cairn terrier, cocker spaniel, Boxer, Maltese, miniature pinscher, miniature schnauzer, Pembroke Welsh corgi, Pomeranian, Portuguese water dog, pug, toy poodle, and Welsh terrier. Median body weight was 8.7 (range, 2.9-43.5) kilograms.

Pre-operative (within 7 days of cholecystectomy) selective hematological and biochemical data is shown in Table 3. Additionally, 7/43 (16%) dogs with a CBC available for review had band neutrophils present. All dogs were administered antibiotics within the perioperative period. Thirty-eight dogs (73%) were administered ampicillin-sulbactam alone, 6 dogs were administered enrofloxacin alone (12%), and 2 dogs (4%) were administered cefazolin alone. The remainder of the dogs received a combination of antibiotics. Three dogs (5%) received ampicillin-sulbactam and metronidazole, and 1 dog each (2%) received ampicillin-sulbactam and cefazolin, ampicillin-sulbactam and enrofloxacin, and ampicillin-sulbactam, metronidazole, and enrofloxacin. At the time of surgery GB rupture was noted in 7/52 (13%, 95% CI [7%, 25%]) dogs. Bile culture was performed in 50/52 (96%, 95% CI [87%, 99%]) dogs. No growth was reported in 44/50 (88%, 95% CI [76%, 94%]). Three dogs had growth of Enterococcus sp., and 1 dog each had growth of Klebsiella sp., Escherichia coli, and Streptococcus sp.

Fifty-one out of fifty-two (98%) dogs had at least 1 hepatic histologic abnormality. Masson's trichrome and rhodanine staining could not be performed for 1 dog, thus these analyses included 51 dogs. The number of dogs with each histologic abnormality is shown in Figure 3. Hepatic fibrosis (37/51; 73%, 95% CI [59%, 83%]), biliary hyperplasia (29/52; 56%, 95% CI [42%, 68%]), and portal inflammation (25/52; 48%, 95% CI [35%, 61%]) were most common (Figures 4 and 5). The most common fibrosis score was F2 (21/51; 41%, 95% CI [29%, 55%]), followed by F0 (14/51; 27%, 95% CI [17%, 41%]), and F3 (9/51; 18%, 95% CI [10%, 30%]). Six dogs (12%, 95% CI [6%, 23%]) had a fibrosis score of F1, and 1 dog (2%, 95% CI [0.4%, 10%]) had a score of F4. Portal inflammation was common (25/52; 48%, 95% CI [35%, 61%]) and mild in all cases. Portal inflammation was solely neutrophilic in 6/25 (24%, 95% CI [12%, 43%]) cases, and mixed neutrophilic and lymphoplasmacytic in 19/25 (76%, 95% CI [57%, 89%]) cases. Few pigmented macrophages were also occasionally seen. In 1 case rare eosinophils were present in addition to the mixed neutrophilic and lymphoplasmacytic infiltrate. Fourteen out of 22 (64%, 95% CI [43%, 80%]) dogs with cholestasis had both extrahepatic (Figure 6) and intrahepatic (Figure 7) cholestasis, whereas 6 had extrahepatic alone (27%, 95% CI [13%, 48%]) and 2 (9%, 95% CI [3%, 28%]) had intrahepatic cholestasis alone in the sections examined. Vacuolar hepatopathy was observed in 22/52 dogs (42%, 95% CI [32%, 56%]; Figure 8). Forty-eight dogs had an interface hepatitis, formerly referred to as piecemeal necrosis, score of 0 (PMN0), and 4 dogs had a score of 1 (PMN1). Thirty-seven dogs had a lobular necrosis score of 0 (LN0), 12 dogs had a lobular necrosis score of 1 (LN1), and 3 dogs had a lobular necrosis score of 2 (LN2). Thirty-six dogs had an activity score of 0 (A0), 13 dogs had an activity score of 1 (A1), and 3 dogs had an activity score of 2 (A2). Thirteen dogs had cholangitis or cholangiohepatitis; 6 had cholangitis alone, whereas the remaining 7 had involvement of the hepatic parenchyma in addition to cholangitis, consistent with cholangiohepatitis. All 13 cases had neutrophilic inflammation (Figure 9) with the exception of 1 where lymphocytes and plasma cells were also described in the portal tracts and adjacent hepatic parenchyma. Cholangitis was described as mild in 5 cases and moderate in 1 case. Cholangiohepatitis was noted to be mild in 5 cases and moderate in 2 cases. The most common copper rhodanine stain score was 1 (26/51; 51%, 95% CI [37%, 65%]; Figure 10), followed by 2 (12/51; 24%, 95% CI [13%, 27%]), and 0 (8/51; 16%, 95% CI [7%, 29%]). Four dogs had a score of 3, zero dogs had a score of 4, and 1 dog had a score of 5. Grade 1 cases had minimal amounts of copper in few randomly scattered hepatocytes and the majority of grade 2 cases only had small amounts of copper in either a few centrilobular hepatocytes or within a small number of randomly scattered hepatocytes.

As was necessary for inclusion, all dogs had outcome data available at the 1 month timepoint. Six dogs were lost to follow up after 1 month so data were available for 46/52 (88%, 95% CI [77%, 95%]) dogs at 3 and 12 months. Ten dogs died in the 12 month period following surgery (10/52; 19%, 95% CI [11%, 32%]). Six were deceased at 1 month. Cause of death noted for these 6 dogs were listed as “ongoing liver issues” for 2 dogs, and unknown for 4 dogs. Three of those 4 dogs passed away naturally at home within 7 days of surgery. An additional dog was deceased by 3 months. Cause of death in this dog was unknown. The remaining 3 dogs died between 3 and 12 months post-cholecystectomy. Cause of death for these 3 additional dogs was unknown for 2 and listed as “general decline” for the third dog. Post-cholecystectomy biochemistry results were available for 4 dogs who were deceased by 12 months. All 4 dogs had persistent elevations in alkaline phosphatase, and 3 dogs had persistent elevations in alanine transferase. Of the 6 dogs who had positive bile bacterial culture results, 1 was deceased at 1 month, a second was deceased at 3 months, and the remaining 4 dogs were alive at 12 months.

Proportions of dogs with all categorical histologic abnormalities alive vs deceased at 1, 3, and 12 months post-cholecystectomy were not different (Table 4). The proportions of dogs alive vs deceased differed depending on the portal fibrosis score at 1 (X² [1, n = 52] = 6.703, P = .01), 3 (X² [1, n = 46] = 5.453, P = .02), and 12 (X² [1, n = 46] = 4.018, P = .04) months post-cholecystectomy (Table 5). Proportions of dogs alive or deceased did not differ based on necroinflammatory scores (LN, PMN, or A0-A3) or rhodanine qualitative copper score at any time point (1, 3, or 12 months) post-cholecystectomy.

The NLR was higher in dogs with hepatic necrosis (P = .006) and cholangitis/cholangiohepatitis (P = .02) when compared to dogs without these lesions. The NLR was not different between dogs with and without portal inflammation (P = .06) or centrilobular hepatitis (P = .09; Figure 11).

4 DISCUSSION

Histologic abnormalities of the liver were common in a group of dogs with GBM. Most abnormalities of the liver were not related to outcome; however, the proportion of dogs alive differed in relation to fibrosis scores at all-time points post-cholecystectomy. This might suggest that chronicity of disease reduces longevity in dogs with GBM. Finally, dogs with GBM and hepatic necrosis and cholangitis/cholangiohepatitis had higher NLR's when compared to dogs without these lesions, therefore the NLR has potential to serve as a biomarker to predict presence of these lesions in dogs with GBM.

The most common histologic abnormalities of the liver in this group of dogs with GBM included hepatic fibrosis, biliary hyperplasia, portal inflammation, cholestasis and vacuolar hepatopathy. Less common abnormalities included centrilobular hepatitis, hepatic necrosis, and cholangitis/cholangiohepatitis. While many previous publications describing GBM in dogs report the histologic, cytologic, and bacteriologic findings related to the GB,^{6-9, 15, 28} comparatively fewer report concurrent histologic abnormalities of the liver.^{11, 12} One previous report found that a high proportion of dogs with GBM have hepatic histologic abnormalities (34/37; 92%) and the most common histologic abnormalities were cholangiohepatitis, biliary hyperplasia, and cholestasis.¹¹ In another study, histologic abnormalities of the liver in 22 dogs with GBM included portal inflammation, hepatic fibrosis, biliary hyperplasia, and vacuolar hepatopathy, although the actual number of individual dogs affected with each lesion is difficult to discern.¹² Hepatic fibrosis was the most common abnormality observed in our cohort. In contrast to earlier studies, we evaluated hepatic fibrosis with Masson's trichrome stain, which may have increased our detection ability. Additionally, since hepatic fibrosis is a response to chronic injury and inflammation in the liver,²⁹ it is possible that our cohort of dogs were affected by GBM for a longer duration of time before cholecystectomy than dogs with GBM in previous studies, however the retrospective nature of the studies makes it challenging to evaluate for this possibility. The prevalence of cholangitis/cholangiohepatitis was similar in our study compared to previous studies.^{11, 12} Consistent with a retrospective review of 54 cases of cholangitis/cholangiohepatitis in dogs, most dogs had a solely neutrophilic infiltrate.³⁰ Bile duct hyperplasia is a characteristic reaction of the liver to various types of insults. In these dogs with GBM, bile duct hyperplasia might have occurred as a consequence of cholestasis, or of portal inflammation and fibrosis.⁵ It is also important to remember that histologic lesions can vary throughout the liver lobes and they might not be present in all examined sections.

We hypothesized that dogs with GBM and hepatitis would be more likely to be deceased at specific timepoints post-cholecystectomy when compared to dogs with GBM without lesions. However, the proportions of dogs alive vs deceased at the 1-, 3-, and 12-month post-cholecystectomy timepoints did not differ in dogs with various forms of hepatitis (cholangitis/cholangiohepatitis, centrilobular hepatitis), or in patients with portal inflammation. Interestingly, Harrison et al reported that cholecystectomized dogs with a histologic diagnosis of cholangitis/cholangiohepatitis had a decreased risk of death when compared to dogs with cholangitis/cholangiohepatitis that did not have cholecystectomy performed.³⁰ Therefore, it is possible that cholecystectomy may help to resolve ongoing cholangitis/cholangiohepatitis in dogs with GBM. However, follow-up hepatic biopsy following cholecystectomy for GBM would be necessary to determine if this is the case. Worsening fibrosis scores may have increased the likelihood that a dog with GBM would be deceased post-cholecystectomy. However, it is important we interpret these results with caution, as we were not able to obtain convincing evidence that death in these cases was related specifically to the presence of fibrosis. Hepatic fibrosis can occur as a consequence of canine chronic hepatitis, lobular dissecting hepatitis, extrahepatic bile duct obstruction, cholangitis, ductal plate abnormalities, right-sided heart disease, and cranial vena cava obstruction in the dog.²⁹ In this group of dogs, the most likely causes of fibrosis include chronic EHBDO or cholangitis, both likely secondary to GBM. While some of the dogs in our study had centrilobular hepatitis, not all the dogs with centrilobular hepatitis had fibrosis.³¹ The suggested impact of hepatic fibrosis in this group of dogs suggest that early cholecystectomy in cases of GBM may lead to improved outcomes, depending on how long it took for fibrosis to develop in these cases. An additional reason to consider early surgical treatment is recent data suggesting outcomes are considerably improved in dogs undergoing cholecystectomy for treatment of GB disease electively compared to non-electively. Mortality rates were 20% for dogs having non-elective surgery, vs 2% for dogs having elective surgery.³² Importantly, because of the small number of dogs in the study as well as the incomplete information regarding cause of euthanasia or death, we cannot conclude from this study that fibrosis is a poor prognostic indicator in cases of GBM. Dogs with GBM may suffer from a variety of complications and concurrent disorders including necrosis and rupture, EHBDO, bacterial infection, cholecystitis, coagulation disturbances, and systemic inflammatory response syndrome,^6-13 all of which may have contributed to death in these patients.

We evaluated the NLR in relation to selected lesions of the liver to examine its utility as a biomarker. We thought it would be particularly useful if the NLR predicted lesions that may impact long-term outcome in dogs with GBM. Dogs with hepatic necrosis and cholangitis/cholangiohepatitis had increased NLR compared to dogs without those lesions. No change in the proportion of dogs alive vs deceased was found in dogs with these lesions post-cholecystectomy. It is important to note, however, that NLR could be increased in dogs with acute pancreatitis, chronic enteropathy, and critical illness.^{17, 19, 21} Thus in this study and future studies its utility to predict certain lesions in dogs with co-morbidities is limited and no cause and effect relationship can be determined.

This study had several important limitations, including collection of retrospective data from multiple institutions over several years, in which approaches to treatment for GBM pre-and post-cholecystectomy may have differed. Furthermore, the effect of any post-operative treatment regimens was not able to be considered when evaluating outcome. Additionally, the sample size was low and sufficient follow-up information was not available to perform survival analysis. Importantly, the reasons for euthanasia or death were not known in the majority of cases, therefore they may have been unrelated to ongoing liver disease, especially considering the median age of dogs in the study. Therefore, the finding that increased fibrosis impacted whether a dog was alive or deceased at specific timepoints must be interpreted with caution. Finally, interobserver agreement is poor when assessing necroinflammatory activity in the liver,³³ therefore it is unknown whether the findings of a different pathologist or team of pathologists would have differed from our team.

In conclusion, hepatic histologic abnormalities are common in dogs with GBM, and advanced fibrosis might be associated with poorer outcomes. These findings might support early surgical intervention in dogs with GBM.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION

Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.