Cholecystitis

Objective: To determine the incidence, diagnostic features, and perioperative predictors of acute cholecystitis after cardiovascular surgery.

Design: Inception cohort study.

Setting: A tertiary care 54-bed cardiothoracic ICU.

Patients: All patients admitted to an ICU after cardiovascular surgery during a 42-month period.

Intervention: Collection of relevant preoperative, operative, and ICU data from a database and medical charts.

Primary outcome: Postoperative acute cholecystitis (AC).

Results: Out of 11,330 admissions, 876 patients stayed in the ICU more than 7 days and 30 of them (3%) developed postoperative AC. AC was diagnosed a median of 26 days after cardiovascular surgery (interquartile range, 11 to 41 days). All patients with AC developed at least two criteria of the systemic inflammatory response syndrome (SIRS), and 16 of them (53%) were vasopressor-dependent on the day of diagnosis. Trends in biochemical testing of liver function were not diagnostic for AC. Death occurred in seven of 17 patients (41%) who underwent cholecystectomy, three of nine patients (33%) treated with percutaneous cholecystostomy, and one of four patients (25%) treated conservatively (p=not significant). Specific earlier predictors of AC were arterial vascular disease, preoperative oxygen delivery less than 430 mL/min [multiplied by] [m.sup.2], longer times on cardiopulmonary bypass, surgical re-exploration, ICU course complicated by cardiac arrhythmia, mechanical ventilation [is greater than or equal to] 3 days, bacteremia, and nosocomial infections.

Conclusion: The incidence of AC is low after cardiovascular surgery. Although SIRS and hemodynamic instability were common at the time of diagnosis, the delayed occurrence and lack of specificity of these features for AC limited their utility for early diagnosis. Specific predictors of AC should be sought in the ICU setting to identify patients who are at risk for AC after cardiovascular surgery. When identified, such predictors can prompt earlier diagnosis and treatment. Further evaluation of the selection criteria for different treatment options is needed in order to decrease the morbidity and mortality associated with AC.

(CHEST 1998; 114: 76-84)

Key words: acute cholecystitis; cardiovascular surgery; morbidity; mortality; predictors

Abbreviations: AC=acute cholecystitis; DNR=do not resuscitate; HIDA=hepatobiliary iminodiacetic acid; ROC=receiver operator curve; SIRS=systemic inflammatory response syndrome

Acute cholecystitis (AC) is an insidious complication that has been increasingly recognized in the critically ill. AC rarely occurs as an isolated event but frequently accompanies multiple organ dysfunction syndrome.[1-3] In 1844, the first reported case of AC was discovered as a fatal complication after treatment of a strangulated femoral hernia? Since then, AC has been described after diverse operative procedures unrelated to the biliary tract, penetrating and blunt trauma, thermal injury, and IV hyperalimentation, as well as in the critically ill.[1,5-12]

Many risk factors have been associated with AC, including hemorrhagic shock, sepsis syndrome, multiple transfusion of blood products, dehydration, refeeding after prolonged fasting, and medication with high doses of opiates.[7,11-13] The precise pathophysiology of AC is still unclear. Visceral hypoperfusion of the gallbladder, increased viscosity and lithogenicity of bile secondary to stasis, endotoxemia, and overproduction of regional inflammatory mediators have been suggested as mechanisms for AC.[9,14,15]

Both early diagnosis and effective treatment are essential to decreasing the morbidity and mortality of AC.[16,17] The diagnosis of AC in ICU patients is frequently delayed because of lack of clinical suspicion.[14] Timely diagnosis is of paramount importance because AC can be a reversible source of sepsis syndrome and acute deterioration of organ function in the ICU.[18] Clinical findings, abnormal blood chemistry, and imaging of the gallbladder can be utilized collectively to improve the diagnostic accuracy of AC in the critically ill.[19-21] Imaging of the gallbladder is commonly performed by ultrasonography, CT scan, or hepatobiliary iminodiacetic acid (HIDA) cholescintigraphy. A high index of clinical suspicion based on known predictors or predisposing factors can enhance early detection of AC. Clinical predictors of AC were addressed previously in a heterogeneous patient case-mix after different operative procedures.[3,5,11,13]

The purpose of this study was to determine the incidence and outcome of AC in a uniform cohort of patients undergoing cardiovascular surgery. The study examined the temporal trends of vital signs, hemodynamics, and diagnostic blood chemistry preceding development of AC. The study also aimed to identify specific predictors that could be ascertained before admission and early in the ICU stay in order to prompt early diagnosis of AC after cardiovascular surgery. The study design was an inception cohort that included all cases of AC and all admissions to the ICU after cardiovascular surgery over a 42-month period.

MATERIALS AND METHODS

Study Population and Data Collection

All patients admitted for cardiovascular surgery between January 1, 1993, and June 30, 1996, formed the study cohort. Among the 11,330 admissions to the ICU, a list of 30 patients with a clinical diagnosis of AC was obtained from a cardiovascular anesthesia database. The cardiovascular anesthesia database included data collected on all patients who had undergone cardiovascular surgery at the institution. Data was collected daily on preprinted forms, concurrent with care, by experienced database personnel who used the medical records, the operation, anesthesia, and perfusion records, and postoperative flow sheets as sources. To confirm the reliability of the data, 10% of the patient charts were randomly selected and reabstracted by physicians blinded to the initial data collection. There was a good agreement between the reference group and the audited sample. The study was approved by the institutional review committee for human research.

Definition of AC

Clinical diagnosis of AC was made based on predefined criteria. The criteria included a combination of clinical features, abnormal laboratory data, and imaging of the gallbladder. Clinical features included fight upper quadrant abdominal pain; tenderness and guarding; palpable mass; signs of diffuse peritoneal irritation; abdominal distention; absent bowel sounds; vomiting, high nasogastric drainage, or ileus; progressive jaundice; unexplained diarrhea; and a core temperature above 38.5 [degrees] C or below 35 [degrees] C. Abnormal laboratory data included WBC count greater than 11x[10.sup.9]/L or less than 4x[10.sup.9]/L; total serum bilirubin [is greater than] 3 mg/dL; aspartate transaminase, alanine transaminase, lactate dehydrogenase, alkaline phosphatase, and gamma glutamyl transpeptidase values greater than 105% of the upper limit of the normal range; and microbiologic study of blood and bile cultures. Ultrasonographic abnormalities indicative of AC included distention of the gallbladder; wall thickness [is greater than] 4 mm; intraluminal biliary sludge; pericholecystic fluid collection; intramural gas bubbles; and a positive Murphy's sign. The presence of at least two ultrasonographic abnormalities on gallbladder imaging was required for the diagnosis of AC. Four patients had incidental findings of gallbladder calculi. On HIDA scintigraphy, the gallbladder was not visualized before and after IV morphine up to 0.1 mg/kg was required for diagnosis of AC.[22] Finally, an inflamed, gangrenous, or perforated gallbladder was noted during laparotomy or diagnostic laparoscopy.

Data and Measurements

Demographic data included age, sex, body mass index, history of diabetes mellitus (insulin- and non-insulin-dependent), arterial vascular disease (previous stroke or carotid artery surgery; history of vascular surgery, transluminal balloon angioplasty, or claudication pain in extremities on physical exertion), hypertension, and smoking. COPD and asthma were documented by history and the patient's need for bronchodilator therapy. A history of congestive heart failure was documented by a history of active symptoms of shortness of breath on exertion, paroxysmal nocturnal dyspnea, orthopnea, or peripheral edema.

Levels of BUN, serum creatinine, albumin, total bilirubin, hemoglobin, and hematocrit were measured during the preoperative work-up. Hemodynamic variables measured or calculated before induction of anesthesia included the heart rate, mean arterial blood pressure, central venous pressure, mean pulmonary artery pressure, cardiac index, stroke volume index, and systemic oxygen delivery. Severe left ventricular dysfunction, which generally corresponds to an ejection fraction of less than 35%, was diagnosed by angiographic ventriculography.

Emergency surgery was performed for one of the following conditions: unstable angina; cardiac shock; ischemic valvular dysfunction that could not be controlled medically; leaking or dissection of thoracic aortic aneurysms; or complications of routine cardiac catheterization or percutaneous transluminal coronary angioplasty. Redo operation implied a prior operation on the heart or thoracic aorta. Surgical procedures performed were classified as coronary artery bypass graft (CABG) isolated or combined with valve surgery, isolated valve repair or replacement surgery, other procedures, and thoracic aorta surgery. Other procedures included placement or removal of automated internal cardiac defibrillation devices, surgical procedures for treatment of arrhythmia, placement of ventricular assist devices, and heart transplantation.

Data on operative events included total cardiopulmonary bypass time, aortic cross-clamp time, number of units of blood products (RBCs and platelets) transfused during the operation, and return to the operating room. In some cases, return to the operating room was required for exploration of persistent postoperative bleeding, tamponade, revision of grafts, or mediastinitis. Arterial blood gases, core body temperature, plasma glucose, infusions of inotropes (amrinone, milrinone, and dobutamine), infusions of vasopressors (dopamine, norepinephrine, epinephrine, and phenylephrine), and infusions of nitrodilators (nitroprusside and nitroglycerine) were recorded after completion of surgery and placement of an intra-aortic balloon pump in the ICU. Biochemical, hematologic, and microbiologic laboratory data from the ICU stay were extracted from the medical records, and the existence of do-not-resuscitate (DNR) orders was recorded. Mortality was defined by death during hospitalization for surgery, regardless of the length of stay, or within 30 days of surgery if the patient had been discharged from hospital.

The medical records of patients with the diagnosis of AC were reviewed for diagnostic procedures, treatment modalities, laboratory data, vital signs, arterial blood gases, and the need for vasopressors and/or inotropes for 7 days preceding the day that AC was diagnosed. The daily maximal body temperature and heart rate, lowest blood pressure, and the number of clinical criteria of systemic inflammatory response syndrome (SIRS) were recorded.[23]

Definitions of Organ Dysfunction

Cardiac dysfunction was defined as a low cardiac output syndrome (by postoperative requirement for inotropes, intraaortic balloon pump, or an open chest for circulatory support) or acute myocardial infarction (by postoperative appearance of new Q waves that were [is greater than or equal to] 40 ms long and [is greater than or equal to] 25% of R wave plus a creatine phosphokinase MB [is greater than or equal to] 50 IU or an aspartate aminotransferase [is greater than or equal to] 80 U/L). Postoperative cardiac arrhythmia included bradyarrhythmia or asystole requiring pacing; supraventricular and ventricular tachycardia; or atrial and ventricular fibrillation.

Pulmonary dysfunction was defined as Pa[O.sub.2]/fraction of inspired oxygen ratio of 150 mm Hg or less and a need for mechanical ventilation. Protracted weaning from ventilator support was recorded if the duration of mechanical ventilation exceeded 3 days.

Renal dysfunction was defined as a postoperative serum creatinine of 3.8 mg/dL, a doubling of the serum creatinine if the preoperative value was greater than 1.9 mg/dL, or a need for renal replacement therapy.

GI dysfunction was defined as small bowel ileus (multiple distended fluid-filled small bowel loops on plain radiographs) beyond the second postoperative day; intolerance of enteral feeding because of persistent vomiting, high volume of gastric aspirate, or abdominal distention necessitating postoperative parenteral nutrition; GI bleeding (indicated by a coffee ground aspirate, melena, or nasogastric aspirate or stool positive for occult blood associated with a drop in hematocrit of 3%); acute pancreatitis; or AC.

Hepatic dysfunction was defined as a rise of total serum bilirubin to 3.0 mg/dL or higher or a doubling of serum bilirubin if the preoperative value was greater than 2.0 mg/dL.

Coagulopathy was defined as a platelet count of less than 50x[10.sup.9]/L. Nosocomial infection was defined as a documented source of infection (lungs, vascular catheters, wound, mediastinum, urinary tract, bloodstream), a positive microbiologic culture, and appropriate clinical features, including a temperature above 38.5 [degrees] C or below 35 [degrees] C, and a WBC count greater than 11 x [10.sup.9] or less than 4 x [10.sup.9]/L.

Neurologic dysfunction was defined as the new onset of postoperative seizures, a focal brain lesion confirmed by clinical findings and/or CT scan; diffuse encephalopathy with severely altered mental status lasting more than 24 h; or failure to awaken postoperatively. Metabolic dysfunction was defined as blood glucose [is greater than or equal to] 300 mg/dL and requirement for administration of insulin.

Cardiac, pulmonary, renal, hepatic, and neurologic dysfunction and coagulopathy were added to give the number of organs exhibiting dysfunction.

Statistical Analysis

All continuous variables were presented as median and inter-quartile range and analyzed by Student's t test or Wilcoxon's rank sum test when appropriate. Variables measured over time were examined by repeated measure analysis of variance. Univariate comparisons of days were performed with a nonparametric test of the median (number of points above median). Categorical variables were expressed as actual numbers as well as a percentage and compared using [chi square] or Fisher's exact test. Variables that had significant association with AC to a value of p [is less than or equal to] 0.1 were entered into a stepwise multiple logistic regression and the computed model was tested for significance and goodness of fit. The discrimination characteristics of predictors for AC before admission and early in the ICU stay were also examined using receiver operator curves (ROCs). An ROC was constructed by plotting the sensitivity vs 1-specificity of the predictors for AC identified from multiple logistic regression. The area under the ROC was proportional to the accuracy or predictive power of the selected predictors to discriminate individual patients in terms of the development of AC. All statistical tests were two-tailed and significance was accepted at p [is less than] 0.05. Statistical analysis was performed using JMP Statistical software, version 3.5.1 (SAS Institute Inc; Cary, NC).

RESULTS

Over a 42-month period, 11,330 patients were admitted after cardiovascular surgery, 876 of whom stayed in the ICU at least 7 days. There were 30 patients with AC, all of whom stayed in the ICU for more than 7 days after cardiovascular surgery. There was no obvious past history of gallbladder disease in this group, even though four patients had incidental gallbladder calculi. During the study period, there were no cases of AC reported in those patients who had a total ICU stay of less than 7 days after cardiovascular surgery. The median number of days before diagnosis of AC was 26 days (25th percentile, 11 days; 75th percentile, 41 days) from the date of cardiovascular surgery. The diagnostic criteria were two or more clinical features in 24 of 30 AC patients (80%), and two or more abnormalities on ultrasonography or CT imaging of the gallbladder in 26 out of 26 patients (100%) who underwent imaging studies. Cholescintigraphy was performed in five patients and confirmed AC in four of them (80%). In four patients, the diagnosis was first established at laparotomy performed for suspicion of intra-abdominal sepsis. Leukocytosis or leukopenia were present in 26 patients with AC (87%). Abnormal results of liver function biochemical tests were seen in 22 patients with AC (73%). Blood cultures were positive in three of 22 patients (17%) and bile culture was positive in seven of 25 patients (28%).

Seventeen patients were treated with open cholecystectomy; six patients were gangrenous and two had free wall perforation of the gallbladder. Nine patients underwent percutaneous cholecystostomy tube drainage as the curative treatment and did not require delayed cholecystectomy. Four patients were treated conservatively with systemic antibiotics and rest of the GI tract. IV hyperalimentation was required in eight patients before AC was diagnosed. Seven patients (41%) died after cholecystectomy, three (33%) died after percutaneous cholecystostomy, and one patient (25%) died after conservative treatment (p=not significant). Delayed diagnosis and treatment of AC was a direct cause of death in three of the nonsurvivors (27%). The remaining nonsurvivors died from multiple organ dysfunction and sepsis that were temporally unrelated to the diagnosis or treatment of AC.

Table 1 illustrates the temporal trend of vital signs and blood chemistry during a 7-day period preceding the diagnosis of AC. A rise in body temperature, heart rate, and WBC count, a decrease in mean arterial blood pressure, and mild metabolic acidosis with compensatory arterial hypocapnia were observed in these patients. On the day of diagnosis, or day 0, all patients (100%) had at least two or more SIRS criteria, 17 patients (57%) had three or more criteria, and nine patients (30%) had four criteria (Table 1). Significant hypotension and vasopressor dependency was also noted in 16 patients (53%) on day 0. Low serum albumin was observed in all patients for 7 days preceding the diagnosis of AC. There were no significant trends in serum bilirubin, liver enzymes, pancreatic enzymes, plasma glucose, or coagulation screen over the 7-day period before AC was diagnosed (Table 1).

Table 1--Temporal Trends of Vital Signs, Blood Chemistry, and Hemodynamic Stability in Acute Cholecystitis After Cardiovascular Surgery(*)

Perioperative factors associated with AC after cardiovascular surgery are reviewed in Table 2. The incidence of diabetes mellitus (22% [n=2,439] in controls vs 37% [n=11] in AC patients, p=0.06), hypertension (53% [n=6,017] in controls vs 53% [n= 16] in AC patients, p= 1.0), history of smoking (60% [n=6,748] in controls vs 56% [n=17] in AC patients, p=0.7), COPD (7% [n=762] in controls vs 10% [n=3] in AC patients, p=0.4), and congestive heart failure (25% [n=2,793] in controls vs 37% [n=11] in AC patients, p=0.2) were similar in the control and AC groups. Arterial vascular disease, intra-aortic balloon pump, inpatient status prior to surgery, elevated BUN, low serum albumin, slightly elevated heart rate, significant elevation of mean pulmonary artery pressure and decreased cardiac index, stroke volume index, hematocrit, and systemic oxygen delivery before cardiovascular surgery were characteristic of the AC group (Table 2).

Emergency surgery (6% [n=689] in controls vs 10% [n=3] in AC patients, p=0.5), isolated CABG (57% [n=6,480] in controls vs 47% [n=14] in AC patients, p=0.3), isolated valve surgery (19% [n=2,170] in controls vs 20% [n=6] in AC patients, p=0.9), and other procedures (12% [n=1,368] in controls vs 10% [n=3] in AC patients, p=0.2) and thoracic aorta surgery (5% [n-518] in controls vs 0% [n=0] in AC patients, p=0.6) were similar in both control and AC groups. Combined CABG and valve surgery was more frequent in the AC group than control group (23% [n=1,299] in controls vs 11% [n=7] in AC patients, p=0.03). Redo operations, long operating and cardiopulmonary bypass time, transfusion of blood products, and return to operating room for surgical re-exploration were also significant operative factors in the AC group (Table 2). Arterial blood gases and core body temperature after surgery and on ICU admission were similar in the control and AC groups. A higher plasma glucose level and more frequent need for inotropes, vasopressors immediately after surgery, and placement of intra-aortic balloon pump in the ICU were significant in the AC group (Table 2). The incidence of cardiac dysfunction, cardiac arrhythmia, gastrointestinal, hepatic, renal, neurologic, and metabolic dysfunction, coagulopathy, bacteremia, and nosocomial infections after cardiovascular surgery were higher in the AC group. The length of initial and total mechanical ventilation, ICU stay, and hospital stay were longer in patients with AC. There were more acute organ dysfunction, DNR orders, and deaths in the AC group.

Multivariate analysis of predictors of AC after cardiovascular surgery are summarized in Table 3. Arterial vascular disease, oxygen delivery below 430 mL/min [multiplied by] [m.sup.2] before cardiovascular surgery, duration on cardiopulmonary bypass, return to operating room for surgical re-exploration, postoperative cardiac arrhythmia, mechanical ventilation for more than 3 days, bacteremia, and nosocomial infections were significant predictors of AC (Hosmer-Lemeshow statistics, p=0.80). These specific predictors as well as the number of organs with dysfunction developing after cardiovascular surgery had excellent discrimination between individual patients for development of AC after surgery (Table 4).

[TABULAR DATA 3 NOT REPRODUCIBLE IN ASCII]

Table 4--The Discrimination Characteristics of Predictors of Acute Cholecystitis After Cardiovascular Surgery(*)

(*) Values shown are median of number of predictors (interquartile range) in each group. Predictors before admission to ICU: arterial vascular disease, oxygen delivery [is less than] 430 mL/min [multiplied by] [m.sup.2], total cardiopulmonary bypass time [greater than] 120 min, and return to operating room for surgical reexploration. Predictors during ICU stay: cardiac arrhythmia, mechanical ventilation for [is less than] 3 days, bacteremia, and nosocomial infection.

([dagger]) Receiver operator curve (area under the curve) expressed as mean [+ or -] SE.

([double dagger]) p<0.0001.

DISCUSSION

The current study identified a 0.3% incidence of AC after cardiovascular surgery. The incidence was 3% among those patients who required prolonged stay in ICU (more than a week). Patients who had an incidental finding of calculi (13%) in the gallbladder were included with acalculous AC in this study, since the same underlying pathophysiologic mechanisms existed in both types of AC in the critically ill.[24] The incidence of cholecystectomy before cardiovascular surgery was less than 2% in the current study, and this did not significantly influence the incidence of postoperative AC. The overall low incidence could reflect the rarity of this condition in the study cohort or, alternatively, only the clinically severe form of AC was identifiable after cardiovascular surgery. Milder forms of AC after cardiovascular surgery could have escaped diagnosis because of lack of clinical manifestation. A prospective surveillance study using strict criteria of imaging of the gallbladder by ultrasonography estimated the incidence of AC at 18% in a cohort of young trauma patients.[25] Many of these cases did not have sufficient clinical manifestations for suspicion of AC. Because we did not rely on gallbladder imaging alone and did not screen all admissions to the ICU for the diagnosis, the incidence of AC after cardiovascular surgery could be underreported. The adjunct use of clinical manifestations and abnormal gallbladder imaging defined a subset of patients with a more severe form of AC that necessitated some form of definitive intervention to avert an adverse outcome.[26] Although strict criteria for gallbladder imaging constituted the mainstay of diagnosis, the coexistence of relevant clinical features still defined the need for therapeutic intervention.[27]

The salient clinical features of AC included new-onset SIRS and hemodynamic instability temporally unrelated to initial cardiovascular surgery. Less than one third of patients had positive microbiologic evidence of infection from either the bloodstream or bile culture at the time of diagnosis, satisfying the definition of true sepsis syndrome.[23] New onset of manifestations of SIRS and hemodynamic instability requiring vasopressor support were only apparent on the day of diagnosis of AC. The temporal trends of vital signs or laboratory data during the preceding 7 days were not helpful in forecasting the development of AC after cardiovascular surgery. Although the occurrence of SIRS, systemic hypotension, and vasopressor dependency were common on the day of diagnosis in patients with AC, these features appeared rather late during the course of illness and were also nonspecific for the diagnosis. SIRS, hypotension, and vasopressor dependency were consistent with a hyperdynamic circulatory derangement in AC, which was also commonly observed in patients with other sources of sepsis.[28]

Specific predictors that can be easily identified before admission and during the initial ICU stay were evaluated in this study to prompt early identification of patients at risk of developing AC after cardiovascular surgery. Although advanced age, sex, diabetes mellitus, and IV hyperalimentation have been reported to be predisposing factors for AC, these factors were found to be of lesser importance after cardiovascular surgery.[3,5,9,12] As in previous studies, arterial vascular disease, multiple surgical procedures, mechanical ventilation, and nosocomial infections were predictors for subsequent development of AC after cardiovascular surgery.[1,7,8,11,12] Characteristic predictors for AC after cardiovascular surgery were inadequate preoperative oxygen transport, longer times on cardiopulmonary bypass, postoperative bacteremia, and cardiac arrhythmia. The predictors were surrogate markers of tissue hypoxia, substantial production of systemic proinflammatory and anti-inflammatory mediators secondary to surgical trauma, splanchnic hypoperfusion, and bacterial translocation after cardiovascular surgery.[29] Postoperative cardiac arrhythmia implied periods of global and regional hypoperfusion continued early during the ICU stay after initial surgery.

Some specific predictors had good ability to identify individual patients who developed AC after cardiovascular surgery: vascular disease, poor oxygen transport, longer cardiopulmonary bypass time, multiple surgical procedures, ICU course complicated by cardiac arrhythmia, bacteremia, nosocomial infections, and ventilator dependency. These predictors, rather than belated manifestations such as SIRS or hemodynamic instability, can be used for early identification of patients at risk for subsequent development of AC in the ICU after cardiovascular surgery. The patient cohort at risk can then be followed more carefully and closely in the ICU by serial ultrasonography of the gallbladder as a surveillance method for early diagnosis of AC. Early diagnosis of AC may facilitate earlier therapeutic intervention by systemic antibiotics and/or percutaneous drainage procedures, and thus may prevent progression to gangrene or perforation of the gallbladder.

The incidence of multiple organ dysfunction was strongly linked to the development of AC, perhaps reflecting an underlying common pathophysiology and etiologic mechanisms after cardiovascular surgery.[29] Although results of liver function tests were abnormal in three fourths of AC cases, there were no acute trends in serum bilirubin, albumin, hepatic transaminase, biliary tract enzymes, or pancreatic enzymes to indicate the evolution of AC after cardiovascular surgery. Biochemical abnormalities of the hepatic, pancreatic and coagulation systems unrelated to AC were perhaps markers of underlying multiple organ dysfunction and splanchnic hypoperfusion after cardiovascular surgery. In fact, dysfunction of three or more organs after cardiovascular surgery was a confounding factor for AC, as shown in other studies.[1-3,8] Although previous studies utilized biochemical abnormalities on liver function tests for the diagnosis of AC, this study did not find such biochemical abnormalities helpful in establishing the diagnosis after cardiovascular surgery.

Cholecystectomy (either by laparotomy or laparoscopy) has long been established as the mainstay of treatment for AC in the critically ill.[8,13,16,28] At cholecystectomy, eight of 17 patients (47%) were noted to have gangrene or perforation of the gallbladder wall as a complication of AC. Previous studies indicated that cholecystectomy was the most effective curative treatment of AC, especially because the incidence of gangrenous necrosis or free wall perforation of gallbladder has been estimated at 40% to 60%.[14,17,24,27] Less aggressive methods of treatment, such as conservative management with systemic antibiotics and gut rest or percutaneous cholecystostomy drainage, have also gained acceptance for the treatment of AC without gangrene or perforation of the gallbladder.[18,25,30-32] In this study, there was no observed difference in mortality among the three treatment options for AC. Preference among the different treatment methods could not be evaluated because the study design was inappropriate, and the sample size was too small to examine each treatment's effectiveness in curing AC without increasing morbidity or mortality. A prospective study in which patients who developed AC are randomized to receive the three different treatment regimens would answer that particular question.

Although conservative management or percutaneous drainage may avoid unnecessary major surgical procedures in critically ill patients, they can also delay the removal of a gangrenous or perforated gallbladder. AC that is complicated by gangrene or perforation was not cured by either conservative management or percutaneous drainage.[14,17,18,24] However, conservative management and a percutaneous drainage procedure have been found to be successful for the treatment of a significant proportion of uncomplicated AC cases, as confirmed by this study.[1,7,10,18,25] Clinical features and laboratory data could not differentiate between complicated and uncomplicated AC cases or help with the decision on the optimal treatment method.[24,32,33] Further studies are needed to define the selection criteria of AC suitable for each treatment modality without compromising patient outcome.

Death was directly attributed to a delay in diagnosis and treatment of AC in 27% of the nonsurvivors. Delay in diagnosis and treatment is still a significant and preventable component of morbidity and mortality associated with AC.[17] A high index of clinical suspicion and treatment tailored to the severity of AC should eliminate some of the adverse outcomes associated with AC. In the majority of patients who developed AC after cardiovascular surgery, AC did not have a direct effect on clinical outcome. All the patients with AC were already seriously ill and suffered multiple organ dysfunction, nosocomial infections, and sepsis refractory to aggressive medical intervention over a long period of hospitalization. The poor response to medical intervention in this group of patients explained the higher frequency of DNR orders and death.

Conclusions

The incidence of AC is uncommon after cardiovascular surgery. Clinically, SIRS and hemodynamic instability were common at the time of diagnosis, but the delayed occurrence of these manifestations limits their utility for early diagnosis of AC. In patients who have undergone cardiovascular surgery, the identification of specific predictors for AC should be attempted early during the ICU stay in order to identify at-risk patients and prompt earlier diagnosis and treatment. Future evaluation of selection criteria for treatment options (conservative management, percutaneous drainage, or cholecystectomy) are needed in order to decrease the morbidity and mortality associated with AC.

REFERENCES

[1] Ouriel K, Green RM, Ricotta JJ, et al. Acute acalculous cholecystitis complicating abdominal aortic aneurysm resection. J Vasc Surg 1984; 1:646-48

[2] Fox MS, Wilk PJ, Weissmann HS, et al. Acute acalculous cholecystitis. Surg Gynecol Obstet 1984; 159:13-16

[3] Hagino RT, Valentine RJ, Clagett GP. Acalculous cholecystitis after aortic reconstruction. J Am Coll Surg 1997; 184: 245-48

[4] Duncan J. Femoral hernia: gangrene of the gallbladder, extravasation of bile; peritonitis; death. North J Med 1844; 2:151-53

[5] Devine RM, Farnell MB, Mucha P. Acute cholecystitis as a complication in surgical patients. Arch Surg 1984; 119:1389-93

[6] Lindberg EF, Brinnan GLB, Smith L. Acalculous cholecystitis in Viet Nam casualties. Ann Surg 1970; 171:152-57

[7] Flancbaum L, Majerus TC, Cox EF. Acute posttraumatic acalculous cholecystitis. Am J Surg 1985; 150:252-56

[8] DuPriest RW, Khaneja SC, Cowley RA. Acute cholecystitis complicating trauma. Ann Surg 1978; 189:84-89

[9] McDermott MW, Scudamore CH, Boileau LO, et al. Acalculous cholecystitis: its role as a complication of major burn injury. Can J Surg 1985; 28:529-33

[10] Hwang TL, Chen MF. Percutaneous gallbladder drainage for acute acalculous cholecystitis during toted parenteral nutrition. Br J Surg 1992; 79:237-38

[11] Orlando R, Gleason E, Drezner AD. Acute acalculous cholecystitis in the critically ill patient. Am J Surg 1983; 145:476

[12] Long TN, Heimbach DM, Carrico CJ. Acalculous cholecystitis in critically ill patients. Am J Surg 1978; 136:31-36

[13] Glenn F. Acute acalculous cholecystitis. Ann Surg 1979; 189:458-65

[14] Glenn F, Becker CG. Acute acalculous cholecystitis, an increasing entity. Ann Surg 1982; 195:131-36

[15] Sakurai T, Ichimiya H, Miyazaki H, et al. Profiling of eicosanoids in inflamed gall bladder wall by gas chromatography with selected-ion monitoring. J Chromatogr 1991; 571:1-18

[16] Herlin P, Ericsson M, Holmin T, et al. Acute acalculous cholecystitis following trauma. Br J Surg 1982; 69:475-76

[17] Johnson LB. The importance of early diagnosis of acute acalculous cholecystitis. Surg Gynecol Obstet 1987; 164:197-203

[18] Lee MJ, Saini S, Brink JA, et al. Treatment of critically ill patients with sepsis of unknown cause: value of percutaneous cholecystostomy. AJR Am J Roentgenol 1991; 156: 1163-66

[19] Mirvis SE, Vainright JR, Nelson AW, et al. The diagnosis of acute acalculous cholecystitis: a comparison of sonography, scintigraphy, and CT. AJR Am J Roentgenol 1986; 147: 1171-75

[20] Cornwell EE, Rodriguez A, Mirvis SE, et al. Acute acalculous cholecystitis in critically injured patients. Ann Surg 1989; 210:52-55

[21] Jennings W, Drabek GA, Miller KA. Significance of sludge and thickened wall in ultrasound evaluation of the gallbladder. Surg Gynecol Obstet 1992; 174:394-98

[22] Flancbaum L, Choban PS. Use of morphine cholescintigraphy in the diagnosis of acute cholecystitis in critically ill patients. Intensive Care Med 1995; 21:120-24

[23] Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest 1992; 101:1644-55

[24] Shapiro MJ, Luchtefeld WB, Kurzweil S, et al. Acute acalculous cholecystitis in the critically ill. Am Surg 1994; 60:335-39

[25] Raunest J, Imhof M, Rauen U, et al; Acute cholecystitis: a complication in severely injured intensive care patients. J Trauma 1992; 32:433-40

[26] Molenat F, Boussuges A, Valantin V, et al. Gallbladder abnormalities in medical ICU patients: an ultrasonographic study. Intensive Care Med 1996; 22:356-58

[27] Coelho JCU, Campos ACL, Moreira M, et al. Acute acalculous cholecystitis. Int Surg 1991; 76:146-48

[28] Savino JA, Scalea TM, Del Guercio LRM. Factors encouraging laparotomy in acalculous cholecystitis. Crit Care Med 1985; 13:377-80

[29] Casey LC: Role of cytokines in the pathogenesis of cardiopulmonary induced multi-system organ dysfunction. Ann Thorac Surg 1993; 56:S92-96

[30] Frazee RC, Nagorney DM, Mucha P. Acute acalculous cholecystitis. Mayo Clin Proc 1989; 64:163-67

[31] Hamy A, Visset J, Likholatnikov D, et al. Percutaneous cholecystostomy for acute cholecystitis in critically ill patients. Surgery 1997; 121:398-401

[32] Rice J, Williams HC, Flint LM, et al. Posttraumatic acalculous cholecystitis. South Med J 1980; 73:14-17

[33] Vauthey JN, Lerut J, Martini M, et al. Indications and limitations of percutaneous cholecystostomy for acute cholecystitis. Surg Gynecol Obstet 1993; 176:49-54

(*) From the Department of Cardiothoracic Anesthesia Division of Anesthesiology and Critical Care Medicine, Cleveland Clinic Foundation, Cleveland, Ohio.

Mohamed Y. Rady MD, PhD, Currently at the Department of Critical Care Medicine, Mayo Clinic Scottsdale, Arizona.

Manuscript received July 31, 1997; revision accepted November 7, 1997.

Reprint requests: Mohamed Y. Rady MD PhD, Department of Critical Care Medicine, Mayo Clinic Scottsdale, 13400 East Shea Blvd, Scottsdale, AZ 85259

COPYRIGHT 1998 American College of Chest Physicians
COPYRIGHT 2000 Gale Group