Microbiologic Evaluation of Gallbladder Bile of Healthy Dogs and Dogs with Iatrogenic Hypercortisolism: A Pilot Study
Parts of this study have been presented at the 17th ECVIM-CA Congress 2007 in Budapest, Hungary.
Corresponding author: Dr P. H. Kook, Dipl. ACVIM, Clinic for Small Animal Internal Medicine, Vetsuisse Faculty, University of Zurich, Winterthurerstr. 260, CH-8057 Zurich, Switzerland, e-mail: firstname.lastname@example.org.
Background: In people, hypercortisolism (HC) has been associated with acalculous cholecystitis and biliary dyskinesia, which may potentiate ascending biliary infections. In dogs, an association between HC and gallbladder disease recently has been documented, although the role of bacteria remains controversial. Furthermore, there is no information on the gallbladder bile microbial flora in healthy dogs.
Objectives: To investigate the microbial flora in gallbladder bile in healthy dogs, the relationship between iatrogenic hyperadrenocorticism and bactibilia and possible changes in biliary microbial flora after cortisol withdrawal in dogs.
Animals: Six control dogs and 6 dogs treated with hydrocortisone.
Methods: Gallbladder bile obtained by percutaneous ultrasound-guided cholecystocentesis was cultured aerobically and anaerobically and examined cytologically before (d0), during (d28, d56, d84), and after (d28p, d56p, d84p) administration of hydrocortisone (8 mg/kg PO q12h).
Results: In the control group, 2/42 bile cultures yielded bacterial growth (Enterococcus sp.; Escherichia coli on d0) and 1/42 bile smears had cytological evidence of bacteria (d28). In the HC group, 2/42 bile cultures yielded bacterial growth (Enterococcus sp. on d28; Bacillus sp. on d28p) and 3/42 bile smears had cytological evidence of bacteria (d84, d84, d28p). All dogs remained healthy throughout the study period (168d).
Conclusions and Clinical Importance: Based on the results of conventional bacterial culture techniques, gallbladder bile of healthy dogs periodically may harbor bacteria, which do not appear to be clinically relevant. A 3-month period of iatrogenic HC was not associated with bactibilia. A higher prevalence of bactibilia may be detected with micromolecular techniques.
percutaneous ultrasound-guided cholecystocentesis
The gallbladder and bile normally are sterile in humans.1–5 Comparable bacteriological studies in veterinary medicine are scarce; results of 1 study suggested that gallbladder bile is sterile in healthy cats.6 Although the existence of biliary microbial flora has not been investigated in dogs, contradictory hypotheses have been proposed: It has been postulated that enteric bacteria commonly reach the liver and are partially excreted in bile and that bile generally is sterile in normal dogs.7 In another study, various bacterial organisms were cultured from liver biopsies in a large percentage of healthy dogs.8
In principle, bacteria can invade the biliary tract by ascending from the duodenum or by the hematogenous route from the hepatic portal venous blood. In humans, liver sinusoids may be exposed periodically to gut-derived bacteria during the course of enterohepatic circulation, and portal bacteremia has been documented experimentally in healthy dogs.9–13 Bacterial invasion of the hepatobiliary system is prevented by potent immunological defense mechanisms (Kupffer cells and biliary IgA) as well as anatomical (tight junctions), physical (bile flow, mucus), and chemical (bile acids) barriers.14 Although it has been shown that corticosteroids impair the phagocytic activity of the reticuloendothelial system,15–18 it is not known whether survival of infectious agents and spill-over into bile may be afforded by loss of 1 of the normal hepatobiliary protective mechanisms in patients with spontaneous hypercortisolism (HC). Interestingly, HC in people has been associated with a high prevalence of acalculous cholecystitis and biliary dyskinesia with hypotonus of the sphincter of Oddi, which may predispose patients to ascending biliary infections.19 In dogs, recent studies have shown an association between HC and biliary mucoceles,20–22 although the role of bacteria in the initiation of mucus hypersecretion remains unclear. Experimental work suggests that biliary infection stimulates mucin hypersecretion.23 Bacterial enzymes also may play an inflammatory role by deconjugating bile acids, making them more toxic to gallbladder epithelium.24 Results of clinical studies on canine mucoceles are conflicting with positive bacterial cultures reported in 9.1–66.0% of dogs.20,25,26 Information on the frequency of bactibilia in dogs with concurrent steroid excess was not provided. Moreover culture results from clinical studies may underestimate the prevalence of concurrent bactibilia because of frequent pre- and perioperative use of antimicrobial drugs. Therefore, the objectives of this experimental study were to investigate the occurrence of microbial flora in gallbladder bile in healthy dogs, the relationship between iatrogenic HC and biliary infection and possible microbiological changes in bile after cortisol withdrawal.
Materials and Methods
This prospective randomized study was approved by the Cantonal Committee for the Authorization of Animal Experimentation (Canton of Zurich, Zurich, Switzerland).
Twelve 3-year-old Beagle dogs, ranging in weight from 10.4 to 16.6 (median, 12.9) kg, were used. There were 6 intact females and 6 intact males, which were considered to be healthy based on the results of physical examination, CBC, serum biochemistry profile (SBP), urinalysis (UA), urine culture, urine protein-to-creatinine ratio, and indirect blood pressure (BP) measurement. The study dogs were concurrently used in another study examining the effects of hydrocortisone on systemic BP, and initial results have previously been published.27–29,a
HC was induced by administration of hydrocortisone, which is the synthetic glucocorticoid most similar to cortisol. The dosage chosen was based on previous studies in which HC was induced in dogs.30
Dogs were randomly allocated to 2 groups of 6 dogs each. Dogs in the control group received a placebo gelatin capsule PO q12h, whereas dogs in the treatment group received hydrocortisoneb at a median dose of 8.5 (range, 7.5–9.6) mg/kg PO q12h for 84 days (HC group). The caretaker and clinicians were blinded to the daily medication. Dogs underwent a physical examination, and percutaneous ultrasound-guided cholecystocentesis (PUC) was carried out before (d0), and 28 (d28), 56 (d56), and 84 (d84) days after starting treatment, as well as 28 (d28p), 56 (d56p), and 84 (d84p) days after withdrawal of hydrocortisone and placebo. CBC, SBP, UA, and urine culture were performed at all examination points in all dogs throughout the study period. The well being of all dogs was monitored daily by an experienced caretaker who worked closely with 2 of the authors (P.H.K. and S.S.). Body weight, appetite, feces, and activity level were monitored daily, and rectal temperature was measured twice daily after PUC. In this model, induction of HC was confirmed by the occurrence of typical clinical and laboratory changes and results of repeated ACTHc stimulation tests.27,a
Specimen Collection. Dogs were fasted for 18–20 hours and sedated with acepromazined (0.01 mg/kg IM) and buprenorphinee (0.05 μg/kg IM) 1 hour before gallbladder aspiration. PUC was carried out aseptically by a transhepatic approach as described previously.31 Briefly, with the dog in dorsal recumbency, a 22 G, 3.5 in. needle attached to a 12 mL syringe was advanced under continuous ultrasound guidance until the needle tip was clearly identified in the middle of the gallbladder lumen. Bile was aspirated until no more fluid could be withdrawn or the gallbladder lumen was no longer visible. The needle was withdrawn after releasing the negative pressure on the syringe. Bile aspirates were submitted to the microbiology laboratory in a Luer lock-capped syringe within 15 minutes of collection. All dogs were reevaluated ultrasonographically 60–120 minutes after PUC to visualize the gallbladder to identify any pericholecystic fluid.
Aerobic and Anaerobic Culture Conditions. Specimens were first cultured in thioglycollate broth enriched with vitamin K and hemin for 72 hours and examined daily for growth. When growth occurred, subculture was carried out for 48 hours (aerobic) and 72 hours (anaerobic) on trypticase soy agar supplemented with 5% sheep blood. Aerobic cultures were incubated in 5% CO2 at 36°C and examined daily for growth for 2 days. Anaerobic specimens were processed and incubated in an anaerobic chamber. Anaerobic cultures were examined for growth daily for 5 days. Contaminants were assessed on the basis of genus identification and number of colonies. Specifically, isolates known to be low-grade pathogens or normal human or canine skin microbial flora that grew in thioglycollate broth only were considered probable contaminants. Bacterial isolates were identified by standard identification procedures.32
Cytology. Immediately after PUC, 3–5 bile smears were made, air dried, stained with a modified Wright's stain and stored in new boxes for later analysis. All smears were examined by a board-certified anatomic pathologist (P.G.) for the presence of crystals, bacteria, inflammatory cells, and epithelial cells.
Induction of HC
In this model, long-term hydrocortisone administration effectively created HC. All dogs receiving hydrocortisone developed clinical signs (polyuria, polydipsia, no regrowth of clipped hair within 1 month, and thinning of ventral abdominal skin with prominent subcutaneous veins) and laboratory abnormalities (stress leukogram, increased alkaline phosphatase activity, isosthenuria) consistent with cortisol excess. During medication, dogs in the I-HC group had higher baseline plasma cortisol concentrations than dogs in the control group, and lower lymphocyte counts as shown in Table 1.
Table 1. Median values and ranges of basal cortisol concentrations and lymphocyte counts in control dogs and dogs of the hypercortisolism (HC) group at the different time points.
|d0||1.75 (0.3–3.4)||1.1 (0.4–3.4)||1.76 (0.74–2.29)||1.69 (1.04–2.12)|
|d28||1.65 (0.7–4.6)||32.4 (21.5–86.9)||2.28 (1.81–2.82)||0.76 (0–1.33)|
|d56||0.5 (0.1–2.5)||30.65 (13.3–57.6)||2.04 (1.41–2.38)||0.7 (0.2–0.94)|
|d84||0.8 (0.2–7.5)||54 (20.3–128.8)||1.84 (1.09–2.27)||0.4 (0–1.54)|
|d28p||0.65 (0.4–1.7)||0.25 (0.1–1.4)||1.75 (1.7–2.04)||1.33 (0.88–1.58)|
|d56p||0.55 (0.2–1.5)||0.5 (0.1–1.9)||1.84 (0.92–2.64)||1.56 (1.28–2.01)|
|d84p||1.25 (0.3–3)||0.55 (0.1–5.7)||1.65 (0.99–2.59)||1.48 (0.85–2.12)|
PUC was carried out a total of 84 times in 12 dogs. Repetitive PUC was straightforward, and no complications occurred during or after the procedure in any of the dogs. The 3.5 in. needle proved to be long enough for placement of the needle tip inside the gallbladder in all cases. Occasionally, a slight push was required when penetrating the muscular gallbladder wall. The viscosity of the bile did not prohibit aspiration through the narrow inner lumen of the 22 G needle, and the gallbladder was emptied completely during PUC in all dogs. The median volume of bile retrieved was 10.5 (range, 5–22) mL. The needle remained in situ no longer than 90 seconds. Follow-up ultrasound examinations identified no free pericholecystic fluid. The only abnormalities occasionally seen were small (2–4 mm) hyperechoic lesions on the mucosal surface of the gallbladder, which were thought to be small hematomas. During the entire study period, all dogs remained bright, alert, and afebrile. They were released from the study in good health and remained healthy for another 8 months before they entered another experiment done by another institute.
Bile and Urine Bacteriology and Bile Cytology
In the healthy control group, 2/42 bile cultures yielded bacterial growth. Both positive cultures (Escherichia coli and Enterococcus spp.) were found on d0. Cytological examination of bile smears of both dogs identified only a few crystals representing concentrated bile. On d28, the dog with the previous positive Enterococcus spp. culture had cytological evidence of a uniform population of Gram-positive cocci but a negative bacterial culture (Table 2).
Table 2. Results of bile cultures and cytologic bile examinations in control dogs and dogs of the hypercortisolism (HC) group at the different time points.
|d0||2 positive cultures (Escherichia coli; Enterococcus spp.)||—||—||—|
|d28||—||1 positive culture (Enterococcus spp.)||1 × gram+ cocci||—|
|d84||—||—||—||2 × gram+ rods|
|d28p|| ||1 positive culture (Bacillus spp.)||—||1 × gram+ rods|
In the HC group, none of the initial bile samples were positive. During treatment with hydrocortisone, 1/24 bile cultures yielded growth of Enterococcus spp. on d28. After discontinuation of hydrocortisone, Bacillus sp. was cultured on d28p in another dog. Both positive cultures were associated with unremarkable bile smear cytological findings. The dog with growth of Bacillus sp. on d28p had had cytological evidence of Gram-positive rods at the previous examination day (d84), although the culture was negative at that time. In another dog, Gram-positive rods were seen cytologically at 2 consecutive examinations (d84 and d28p) (Table 2). None of the bile smears of the remaining dogs had evidence of crystals, epithelial cells or inflammatory cells. All growth on culture plates was visible after 24–48 hours, and there were no positive subcultures. Results of all 84 urine cultures were negative. Results of routine laboratory analyses (CBC, SBP, UA) indicated no evidence of a systemic inflammatory response or hepatocellular damage. Statistical analysis was not conducted because of the small number of positive cultures and cytological results.
An increase in gallbladder disease associated with HC in dogs prompted us to investigate bactibilia, which has not been examined before. The microbial flora of gallbladder bile in healthy dogs also was investigated and served as a control. Our findings contradict the assumption that bile generally is sterile in healthy dogs. In the control group, 2 of 6 (33%) dogs had characteristic enteric bacteria in cultures of gallbladder bile on 1 of 7 occasions, and cytological evidence of bactibilia was obtained on 1 other occasion. Thus, 3 of 42 (7.1%) bile specimens were not sterile. In fact, by combining the results of HC and control dogs, a total of 5 of 12 (41%) dogs had evidence of bacteria in their bile at some point during the study. This is in contrast to the current assumption that canine bile is sterile, but must be confirmed in a larger group of dogs.
The serial clinical, hematological, biochemical, and ultrasonographic results as well as concomitant urine cultures were normal in this group and therefore, the health of the animals was not affected. In particular, the presence of bacteria was not accompanied by cytological inflammatory changes in the bile. There are 2 plausible explanations for the transient bactibilia found in 2 different dogs during the course of the study: First, transient bactibilia may be a normal phenomenon in healthy dogs that does not affect their health. Because results of experimental and clinical work suggest continuous hepatobiliary exposure to enteric bacteria,8,9,12,13 normal as well as diseased canine livers likely are commonly exposed to enteric bacteria and a few organisms may escape intrinsic killing mechanisms. However, provided that the continuous flushing action of bile and bacteriostatic properties of bile acids are maintained, it is unlikely that cholangitis and cholecystitis will develop.14 Secondly, our findings may represent simple contamination, either during PUC or during transport and cultivation. However, this seems unlikely, because all procedures were carried out under the same aseptic conditions. Furthermore, E. coli and Enterococcus spp. are typical enteric bacteria rather than skin contaminants. In a recent study of 20 healthy dogs, the livers of 12 dogs harbored different potentially pathogenic bacteria, which did not cause detectable disease.8 This also supports the concept that bile may not be constantly sterile, and that positive bacterial cultures were not mere contamination.
Our findings in dogs with HC do not support the idea that transient HC is an important risk factor for bactibilia and that biliary disease associated with HC is triggered by bactibilia. In fact, cytological and bacteriological evidence of bactibilia was found in only 3 of 24 (12.5%) samples from 3 different dogs during HC treatment. The culture of Bacillus spp. after cortisol withdrawal on d28p is more difficult to assess. Although it seems likely that this finding represents skin contaminant33 the same biliary isolate has been found in canine cholelithiais before.34
In several control and HC dogs, there were conflicting bacteriological and cytological results. Four bile samples had cytological evidence of a uniform bacterial population but negative cultures, which indicates that contamination was not likely. It is not clear whether or not this discrepancy points to suboptimal culture conditions. Our culture techniques differed from those used in a recent microbiological study of hepatobiliary disease in dogs and cats, in which longer aerobic and anaerobic incubation times were used.35 However, this aspect is difficult to assess because the percentage of cultures that yielded growth and the associated incubation times were not provided in that study. Further speciation of bacterial genera would have been helpful to evaluate potential contamination, but these techniques were not routinely available at the time the study was performed. Another explanation for the above-mentioned discrepancy may be growth-inhibiting factors, such as lysozyme and lactoferrin, which in people constitute activated local defense mechanisms against bacterial infection.36
Our study had several limitations. The induced HC may not have lasted long enough to mimic spontaneous HC, which has a more insidious and chronic course and may result in bacterial colonization of the gallbladder more frequently because of biliary sphincter dysmotility.19 Interestingly in this context, we also failed to see an increase in urinary tract infections in this population in contrast to another study in dogs on long term glucocorticoid treatment (> 6 months) in which UTI's were more common.37 Furthermore, bacterial products such as lipopolysaccharides in bile samples were not investigated. These substances have been shown to result in mucin hypersecretion in cultured epithelial cells from canine gallbladders.23 It could therefore be argued that a negative biliary culture result may not be sufficient to rule out bactibilia as a potential trigger for gallbladder disease associated with altered bile fluidity in dogs. For this reason and to clarify the ambiguous relationship between cytology and culture results, the use of molecular identification methods, such as 16S rDNA PCR, would have been desirable. The study however was performed at a time when these tools were not routinely available.
Recently, molecular methods, such as 16S rDNA analysis, have been used to study the bacterial ecosystem in various compartments of the GI tract. Several studies have demonstrated that molecular methods are far more sensitive than classical culture-based methods in identifying bacterial species.38,39 However, in a recent study on feline bile, DNA sequencing was only able to identify bacteria in 1 bile sample for which bacteria had not already been identified with bacterial cultures.f Thus, although it appears desirable to expand this study by molecular methods, the latter study suggests that positive samples may not be more frequently detected by molecular methods.
Another limitation was the relatively small number of dogs (12 laboratory Beagles). This study originally was part of another project assessing the effects of HC on systemic arterial BP,27–29,a and the number of dogs used was determined by the other study. Another limitation was the potential risk associated with repeated PUC in hypercortisolemic dogs. Although PUC has been shown to be a safe and straightforward procedure in healthy dogs,31 the risks associated with repeated PUC in hypercortisolemic dogs are unknown. For these reasons, the authors considered the present work a pilot study in order to insure that experimental design is optimized in more definitive studies to minimize the number of animals enrolled.
In summary, conventional culture of gallbladder bile of healthy dogs occasionally resulted in bacterial growth, although this finding did not appear to be clinically relevant. Iatrogenic HC over the course of 3 months was not associated with bactibilia. A higher rate of bactibilia might be identified with micromolecular techniques. In addition to determining the prevalence of bactibilia with more sensitive methods, further studies on a potential relationship between gallbladder disease and HC that specifically investigate biliary tract motility21 are warranted.
aSchellenberg S, Wenger M, Reusch CE, et al. Course of hematological and biochemical changes during and after long-term hydrocortisone treatment in healthy beagles. J Vet Int Med 2008;22:1476 (abstract)
bHotz Pharmacy, Kusnacht, Switzerland
cSynacthen, Novartis Pharma Schweiz AG, Bern, Switzerland
dPrequilan, Fatro, SPA, Ozzano Emilia, Italy
eTemgesic, Essex Pharma GmbH, Germany
fMorgan M, Rankin S, Berent A, et al. Prospective evaluation for bacterial infection in hepatic tissue and bile of cats with diffuse hepatobiliary disease. J Vet Int Med 2008;22:806 (abstract)