Reprints Available Directly from the Publisher Photocopying Permitted by License Only an Analysis of Infectious Failures a Chola G'" Cute N Tls Patients and Methods Patient Selection

To determine the factors responsible for therapeutic failures in acute cholangitis, a series of 127 patients was analyzed. There were 64 females and 63 males whose mean age was 57.2 years. Ninety-four (74.0%) ofthese patients were clinically cured with initial measures, whereas 33 patients (26%) failed initial therapy for an infectious reason. No differences were observed between the two groups in regard to age and gender. However, more patients in the group that failed had a malignant cause for their bile duct obstruction (72.7% vs. 42.6%, p < 0.01) and had a pretreatment positive blood culture (45.5% vs. 13.8%, p < 0.01). Patients who failed had a higher mean total bilirubin level (9.7 mg/dl vs. 5.5 mg/dl, p < 0.005) and more ofthem had a level greater than 2.2 mg/dl (97% vs. 69.9%, p < 0.001). Also, more bile cultures were initially positive (93.9% vs. 76.6%, p < 0.05) and more organisms were isolated per culture (3.88 vs. 2.86, p < 0.03) in the patients who failed. In addition, more patients failed who had two or more organisms in the bile (33% vs. 8.3%, p < 0.02). Patients in whom Candida, or any panresistant organism was isolated also tended to fail. Multivariant analysis showed that malignancy, bacteremia, bilirubin > 2.2 mg/dl, > 2 organisms in the bile and a panresistant organism were the best predictors oftreatment failure with a serum bilirubin level > 2.2 mg/dl being the variable that increases a patient's log-odds ratio offailure the greatest. In conclusion, patients with acute cholangitis who have an increased chance to fail initial therapy can be identified, and treatment altered accordingly. Acute cholangitis is infection in an obstructed biliary system from both benign and malignant causes. Initial therapy is directed toward general support ofthe patient including fluid and electrolyte replacement, correction of metabolic derangements, broad spectrum antibiotic therapy, and biliary drainage. The majority ofpatients with acute cholangitis are cured of their infection by these efforts. However, there are patients in whom this early treatment ofcholangitis fails and who require alteration in therapy, such as a change in antibiotics, further biliary drainage, or treatment of an intervening infection like a wound infection or pneumonia. The factors involved in these infectious failures are not completely clear, and the purpose ofthis study is to analyze a series ofpatients with acute cholangitis to elucidate those elements that play a role in patients who fail initial therapy …


iteria were a clin
cal diagnosis ofacute cholangitis supported by either pain or tenderness in the right upper quadrant orjaundice, and hypothermia of less than 36.5 degrees C.; fever of more than 38.0 degrees C.; septic shock; or white blood cell (WBC) count of more than 10 x 109/L.Patients were not included if another medical or surgical illness could explain the illness.included in this study had obstruction ofthe biliary tract documented roentgenographically and/or at surgery.


Patient Evaluation

Pretreatment blood cultures were obtained from all patients.Aerobic and anaerobic bile cultures were obtained from indwelling biliary tubes at the time ofper- cutaneous draining, or at surgery and were transported to the microbiology laboratory in anaerobic transport media.All antibiotics were given intravenously for a minimum of five days unless a clinical infectious failure occurred at which time the antibiotics were usually changed.A,patient was considered a clinical cure if the signs and symptoms of cholangitis completely abated with no evidence of infection in the follow-up period.A

tient was judged im
roved if there was a signifcant improve- ment in the clinical findings, but not complete resolution of the infection.The relapse category was used for patients who initially improved only to have recurrence ofinfection within three weeks oftherapy.A failure was defined as no demonstrable response to therapy or the development ofa significant postoperative infection in patients undergoing an operation.Abdominal wounds were considered inf - cted ifpurulent material that subsequently grew organisms on culture could be obtained from them.


StatisticalAnalysis

Student's two-tailed t-test and the chi-square test were used to compare the differences between the study groups and treatment results.Logistic regression analysis was employed to evaluate the relative importance of the differences observed between the study groups for the individual variables.The data are expressed as mean (plus or minus standard error of the mean) or as a percent of patients.


Demographics

There were 127 patients with acute cholangitis eligible for analysis.The number of patients in the improved and relapse categories was too small for valid statistical analysis, and

ese patients were in
luded in the clinical cure and failure groups, respectively.Therefore, 94 patients (74%) had a clinical cure and 33 (26%) werejudged to be a clinical failure.The reasons for failure were persistent sepsis in 28 patients (85%) and a wound infection in five patients (15%).

The mean age of the patients in this series was 57.2 years, and there were 64 females (50.4%) and 63 males (49.6%).There were no

ifferences be
ween the cure and failure groups with regard to age or gender.In the cured group the mean age was 56.8 + 1.5 years and there were 47 females (50%) compared to a mean age of 58.4 + 1.9 years and 17 females (51.5%) in the failure group.No difference was observed in the admitting temperature between the two groups (38.5 + 0.08 C in those patients who were cured vs. 38.8+ 0.15 C in those who failed).

Sixty-three patients in this series had a benign cause for their bile duct obs ruction (stones-43, stricture-15, sderos- ing cholangitis -3, other-2), and 64 had a malignant etiology (bile duct cancer 41, pancreatic cancer-11, colon cancer- 4, gallbladder cancer -3, other-5).Nearly three-fourths (24/ 33) of the failure patients had a malignant obstruction compared to only 43% (40/94) of the cured patients (p < 0.01).

The comorbidities present in the patients in this series are listed in Table 1.The hematopoietic disease observed in these patients was generally chronic anemia although one patient in the failure group ha hypersplenism and one in the cured group had polycythemia vera.Patients were considered to have cardiac disease ifthey had underlying coronary artery disease, valvular heart disease or congestive heart failure and hepatic disease if they had documented cirrhosis or hepatitis.The renal disease refers to preexisting chronic renal failure.No differences were noted between the groups as 72 patients (76.6%) in the cured group had a major comorbidity, while 23 patients (69.7%) in the failure group were so afflicted.

Three patients in this series presented in septic shock, 2 (6%) in the group that failed and (1%) in the group that was cured.In addition 3 patients (11%), all in the failure group, had mild acute reversible renal failure associated with their attack of acute cholangitis.The finding was significant (p < 0.03).


Lzoratory Data

The laboratory data is shown in Table 2.The two groups were similar with respect to initial white blood cell count,  23 (69.7%) alkaline phosphatase, SGOT, SGFr, creatinine, and BUN.The only significant difference observed was a higher total bilirubin level in the failure group (9.7 + 1.3 mg/dl vs. 5.5 + 0.5 mg/dl, p < 0.005).In addition, more patients in the group that failed had a total bilirubin level greater than 2.2 mg/dl (97% vs. 69.9%,p <0.001).


Operations andProcedures

In addition to antibiotic treatment, biliary drainage was liberally employed in the patients in this series.Biliary drainage tube

primarily of t
e transhepatic type, were the most common modality employed, and were used to stent anastomoses and palliate malignancies as well as to drain nfected bile.More patients in the failure group had tube drainage compared to those that were cured (90.9% vs. 60.6%,p < 0.001).The vast majority ofthese tubes were already present prior to the attack of acute cholangitis, and treatment included irrigation, replacement, and repositioning.There were only two patients in the failure group in whom complete drainage ofthe biliary tract was probably not accomplished, and both had an unresectable malignancy.

Concerning operations and other types ofinvasive procedures exclusive ofbiliary tube drainage, no differences were noted between the two study groups.In the group that was cured, 50% of patients had such a procedure compared to 51.5% ofthose who failed.The specific pro- cedures are shown in Ta le 3.


Bacteriology

Twenty-eight patients (22%) in this series had a positive blood culture associated with their attack of acute cholangitis, and this bacteremia occurred more frequently in the failure patients (45.5% vs. 13.8%,p < 0.01).In most instances, the same organism found in the blood was

entually reco
ered from the bile.In addition, 103 patients (81.1%) had bactibilia on the initial bile culture,  and the failure patients were more likely than those who were cured to have a positive culture (93.9% vs. 76.6%,p < 0.05).The patients who failed also had more organisms per specimen (3.88 vs. 2.86, p < 0.03), and more of them had two or more organisms in the bile (33% vs. 8.3%, p < 0.02).

The actual organisms cultured from the bile in the two groups of patients are shown in Table 4.There was a tendency for Enterobacter, Streptococcus sp, and Candida to be present more often in the failure patients, but this tendency was significant only for Candida (27.3% vs. 10.6%, p < 0.05).Other types ofyeast were isolated occa- sionally in this series, and all yeast as a group was also present significantly more often in the patients who failed (33% vs. 11.7%,p < 0.02).

No pure cultures ofanaerobic organisms were isolated in this series ofpatients, and there was the same percentage ofmixed aerobic/anaerobic cultures in the two groups (18.2% in the failure patients and 17.0% in the cured pa- tients).Most patients had pure aerobic cultures.Eight patients, each with at least three bacteria in their bile culture, were found to have an organism resistant to all ant biotics tested for.Five of these patients failed, three were cured (15.2% vs. 3.2%, p < 0.05), and all had a malig- nant etiology.Lastly, no particular combinations of Most ofthe patients who failed had repeat bile cultures; and as expected, there was persistence of the prevalent organisms and a tendency toward increasing antibiotic resistance, which was particularly evident with Enterobacter and Pseudomonas.Also there was a tendency to isolate yeast more often in repeat

le cultures, although thi
was not statistically significant (44% vs. 33%).

When patients were grouped by benign or malignant etiology oftheir cholangitis, it was evident that there were major differences in the organisms encountered.Specifically, Pseudomonas sp, Enterobacter sp, Enterococcus sp, Klebsiella sp, and all yeast were encountered signi icantly more frequently in patients with malignancies (p < 0.005 for each).It should be noted that only one patient with a benign cause for cholangitis had yeast on the initial culture (p < 0.00001).


Deaths and Complications

Five patients (4%) in this series died during the hospitalization for acute cholangitis, and each had a malignant bile duct obstruction.No significant difference in mor ality was noted between the patients who were cured and those who failed (3.2% vs. 6.1%).Hepatic failure was a factor in the death of all five patients.

The major complications encountered in this series were transient renal toxicity and hematopoietic toxicity, which was manifest as an elevated prothrombin or partial thromboplastin time.A greater percentage ofpatients in the failure group developed a complication (21.2% vs. 9.6%) as well as

enal (12.1% vs. 6.4%)
r hematopoietic (15.2% vs. 5.3% ), but these differences were not significant.

Five patients in this series had liver abscesses, and there were no differences between the groups, as four of these patients (4.3%) were cured and one (3%) was a failure.failure.In addition, a serum bilirubin level _> 2.2 mg/dl is the one variable that increases an individual patient's logodds ratio of failure the greatest, when all the other variables are corrected for.These log-odds ratios for all five parameters are shown in Table 5.


Multivariant Analysis

Applying the mathematical model for the parameters of malignancy, positive blood culture, bilirubin level >_ 2.2 mg/dl, and multiple ( >_ 2) organisms in the bile produces the predicted probabilitie

for failure
exhibited in Table 6.The presence of panresistant organisms is purposely ex- cluded as this information is generally not available within a time frame to be clinically useful.As shown, if all four parameters are absent, there is a 4% predicated probability of failure, whereas when all four are present, there is a greater than 99% predicted probability offailure.In addi- tion, bilirubin level alone is more predictive offailure than malignancy alone or combined with either positive blood culture or multiple organisms in the bile.


DISCUSSION

The purpose of this review was to identify in a retrospective fashion those factors involved in infectious treatment failures in patients with acute cholangitis.

Of note, three of the five most significant variables identified by multivariant analysis, i.e. positive blood The results reported above suggest that the single variables ofmalignancy, positive blood culture, acute renal failure, bactibilia, multiple organisms in the bile, yeast and panresistant organisms in the bile, elevated bilirubin level, and biliary drainage tubes are risk factors influencing treatment failure in patients with acute cholangitis.Multivariant analysis was performed using logistic regression o analyze the relative significance of these variables as risk factors both independently and in combination.It was found that various combinations of positive blood culture, malignancy, bilirubin _> 2.2 mg/dl, multiple ( 2) organisms in the bile, and panresistant organisms in the bile are the best predictors oftreatment culture, > 2 organisms in the bile, and the presence of a panresistant organism in the bile, are infectious parameters.The white blood cell count, temperature, age and gender, which have been useful predictors of more serious illness or adverse events in some analyses, [4][5][6] were not discriminatory in this study.

A serum bilirubin level greater than 2.2 mg/dl was the factor that increased an individual patient's odds offailure the most.The role ofjaundice or hyperbilirubinemia in increasing postoperative morbidity and mortality in patients undergoing biliary tract surgery has not been com- pletely elucidated..Several reports suggest that an elevated bilirubin is related to the development of postoperative complications and operative mortality, 5-8 and some studies have shown that reducing the bilirubin level preoperatively in patients with obstructivejaundice can reduce the incidence of postoperative complications and death 9,1.Other reports, though, have shown tha preoperative re- duction ofthe bilirubin level has no impact on postopera- tive problems , 12.In addition, it is suggested in one study thatjaundice per se does not contribute substantially to an undesirable postoperative outcome13.In this study, the determinant of poor outcome was an infectious failure following initial therapeutic measures for acute cholangi- tis.It is known that patients with chronic obstructive jaundice have impaired neutrophil functions and other host defense mechanisms such as synthesis ofacute-phase proteins and complement14, and conceivably these deficits were operative in some patients in this series.

A malignant etiology for bile duct obstruction is an- other major variable that increased the chance offailure in this study.There see s to be rather universal agreement that malignancy correlates with poor results in patients with obstructivejaundice who have surgery, with or with- out infection 5-7,13,5.It is also well recognized that both cellular and humoral immunity are impaired either be- cause of the underlying malignancy and/or the antine- oplastic therapy6.In addition, malnutrition is a frequent complication ofmalignancy which further decreases host defense.This level ofimmunocompromise in patients with cancer makes them more susceptible to infectious complications such as acute cholangitis, and makes the treatment ofthese problems all the more difficult.

Bacteremia is a frequent occurrence in patients with acute cholangitis as evidenced by the 22% incidence in our patients.It has been shown that obstructivejaundice can promote bacterial translocation from the gut in an animal model, possibly by an impairment in the mucosal barrier function, as well as by imp ired host immune defenses17.It is not known if this leads to bacteremia or bactibilia or both.The bacteremia may also arise from an obstructed, colonized bile duct.Since dissemination of microorgan-isms into the circulation can result when local host defense mechanisms fail anywhereTM, bacteremia is a natural conse- quence of acute cholangitis in many patients.It is not surprising that bacteremia was a significant predictor of infectious failure in our patients.Dissemination of mi- crobes via the blood stream will activate the systemic inflammatory response which can be a potential host liabil- ity ifmultiorgan failure intervenes8.

Most patients in this series had bactibilia.The organ- isms isolated are similar to those that have been reported by others9-22.Like these reports, multiple organisms were isolated from the bile i the majority ofthe patients in this study.However, those patients in whom the bile cultures grew two or more organisms were at greater risk offailure than those patients whose cultures grew none or, at most, one organism.Since a sterile bile culture could be inter- preted as eradication ofinfection, one would expect these patients to have less infectious sequelae following ad- equate therapy.Moreover, the presence of two or more organisms might be an indication of a greater infectious burden or bacterial synergy, both ofwhich are more dif- ficult to treat, especially in patients who arejaundiced and/ or have an underlying malignancy.

The only organisms that were present significantly more often in the patients who failed were panresistant ones and yeast.In our experience complete eradication of all the organisms in the bile is uncommon, even in those patients who are cured of their acute infectious illness.Complete eradication would be particularly difficult to achieve for a patient with either a anresistant organism or a tenacious opportunistic pathogen such as yeast.A bile culture con- taining a panresistant organism was one ofthe major de- terminants offailure in the multivariant analysis.It should be noted that Enterococcus was present in the bile culture of the failure patients at least as often as the other major pathogens such as E. coli and Klebsiella, but its signifi- cance regarding predilection for failure was not apparent.Whether Enterococcus is a major pathogen in non-nosoco- mial intra-abdominal infection has not been completely 23 elucidated and the results of the current study do not help to clarify this question.

Biliary drainage tubes were present in many ofthe pa- tients in this series and were used for a variety of indica- tions, including drainage of infected bile and stenting of unresectable malignancies or tenuous anastomoses.Many times in this series occlusion or some other malfunction of the tube was the initiating event in the development of an attack of acute cholangitis.Irrigation, repositioning, or even replacement was often required to control the infec- tion.Even though the presence of a biliary tube was not one ofthe most important factors influencing an i fectious failure following initial treatment of acute cholangitis, tubes in general do complicate matters.First, they are a foreign body generally draining to the outside providing easy access offlora into the biliary tract and making com- plete eradication impossible, and second, they precipitate repeat attacks of acute cholangitis in some patients who then receive numerous courses of antibiotics5.In addi- tion, some clinicians place these patients on long-term suppressive oral antibiotic therapy which can lead to the development ofresistant organisms and the emergence of opportunistic pathogens, particularly yeast.

Complications of therapy and death were somewhat more prevalent in those patients who were infectious fail- ure in this series than in those who were cured, but these differences were not significant.One would expect pa- tients with more serious infections to have more problems, but this trend could not be fully documented in this report.


S

A series of patients with acute cholangitis has been analyzed.For those patients who do not respond to initial therapy, several risk factors have been identified.The most important of these are a serum bilirubin greater than 2.2 mg/dl, bacteremia, a malignant cause ofthe bil

duct obstruction, t
o or more organisms in the bile culture, and a panresistant organism in the bile.Most of these parameters are generally known within the first 48 to 72 hours of therapy.Therefore, early in the course of treatment, the clinician should be alerted to the possibility that the patient with two or more of these risk factors has a greater chance to fail, and be prepared to alter the treatment plan.Such alterations should include bro- adening or changing the antibiotic coverage, increased diligence in monitoring for signs ofimpending sepsis, and assuring adequate and complete drainage of the biliary tract.


INVITED COMMENTARY

An Analysis OfInfectious Failures In Acute Cholangitis Acute cholangitis is still associated with a substantial morbidity and mortality despite improvements in modem antibiotic therapy and intensive care treatment.The identification offactors predisposing to treatment failure should thus of great clinical value.The present study demonstrates that high bilirubin levels, positive bile cultures, multiple bacterial growth, positive blood cultures prior to treatment and malignancy correlate with a higher isk oftreatment failure in acute cholangitis.The finding of increased bilirubin level, reported to. the most important factor, could have several explanations.Biliary obstru- ction induces a variety ofalterations in host defense, such as an impairment in RES function particularly Kupffer cell function-3.This could be due to factors like increased intraductal pressure, increased plasma concentrations of antimetabolic substances or depletion of serum opsonins otherwise promoting pha ocytosis.Furthermore, increased bilirubin levels by inhibiting cellular respiratory enzymes might impair energy dependent phagocytosis 4.

The endotoxemia noted in patients with obstructive jaundice probably also has an influence, e.g. by inducing macrophage dysfunction 5,6 and the lack of intestinal luminal bile salts in biliary obstruction also increases endo- toxemia as bile salts possess a direct detergent effect on endotoxin7.The presence ofbiliary obstruction might also constitute a hindrance to hepatocyte excretion and it might thus be more difficult to obtain the required con- centrations ofantibiotics in bile.

The present study demonstrates bactibilia as a factor predisposing to treatment failure.In patients with com- mon bile duct stones and acute suppurative chotangitis, bacterial cultures from bile seem to be always positive 8,9.It is to be emphasized that anaerobic bacteria, mainly Clos- tridium and Bacteroides, with improved culture techniques are demonstrated in maybe 20-30% of positive cultures ..T hus, antibiotics also covering the anaerobic flora should be considered.Otherwise, as also demon- strated in the present study, bacterial culture findings fairly well represent bacteria present in the distant small intestine 12,13.The use of biliary drainage tubes tended, though not significantly, to predispose