AIM: To determine patient and process of care factors associated with performance of timely laparoscopic cholecystectomy for acute cholecystitis.

METHODS: A retrospective medical record review of 88 consecutive patients with acute cholecystitis was conducted. Data collected included demographic data, co-morbidities, symptoms and physical findings at presentation, laboratory and radiological investigations, length of stay, complications, and admission service (medical or surgical). Patients not undergoing cholecystectomy during this hospitalization were excluded from analysis. Hierarchical generalized linear models were constructed to assess the association of pre-operative diagnostic procedures, presenting signs, and admitting service with time to surgery.

RESULTS: Seventy cases met inclusion and exclusion criteria, among which 12 were admitted to the medical service and 58 to the surgical service. Mean ± SD time to surgery was 39.3 ± 43 h, with 87% of operations performed within 72 h of hospital arrival. In the adjusted models, longer time to surgery was associated with number of diagnostic studies and endoscopic retrograde cholangio-pancreatography (ERCP, P = 0.01) as well with admission to medical service without adjustment for ERCP (P < 0.05). Patients undergoing both magnetic resonance cholangiopancreatography (MRCP) and computed tomography (CT) scans experienced the longest waits for surgery. Patients admitted to the surgical versus medical service underwent surgery earlier (30.4 ± 34.9 vs 82.7 ± 55.1 h, P < 0.01), had less post-operative complications (12% vs 58%, P < 0.01), and shorter length of stay (4.3 ± 3.4 vs 8.1 ± 5.2 d, P < 0.01).

CONCLUSION: Admission to the medical service and performance of numerous diagnostic procedures, ERCP, or MRCP combined with CT scan were associated with longer time to surgery. Expeditious performance of ERCP and MRCP and admission of medically stable patients with suspected cholecystitis to the surgical service to speed up time to surgery should be considered.

Nearly 700 000 cholecystectomies are performed in the United States annually, making it the most commonly performed operation on the digestive tract[12]. The optimal timing for the performance of laparoscopic cholecystectomy (LC) in the subset of patients with acute cholecystitis has been a point of much discussion. Initially there was concern about potential dangers of performing an operation on an acutely inflamed gallbladder with instruments that lacked the feedback, dexterity, and precision of human hands[34]. Studies within the past ten years, however, have determined that laparoscopic cholecystectomy for patients with acute cholecystitis within three days of admission is not only safe but also associated with fewer post-operative complications, decreased morbidity, decreased length of hospital stay, and decreased overall cost when compared to delayed or interval LC[5–9]. Cognizance of the general preferability of prompt surgery, however, may not be uniform among all caretakers. Hence, we examined the time to laparoscopic cholecystectomy in our institution. In order to help identify potential areas for improvement in current management, we also examined the association of presenting signs, pre-operative imaging studies and procedures, and admitting service (medicine versus surgery) with time to operation.

We performed a retrospective medical record review of patients identified by a primary or secondary discharge diagnosis of acute cholecystitis (ICD codes 575.0, 575.10, 575.11 and 575.12) at Norwalk Hospital between March 2003 and December 2005. The diagnosis of acute cholecystitis was made at the time of surgery by experienced biliary surgeons and confirmed by acute inflammatory changes being found on pathologic examination[10]. Patients not undergoing cholecystectomy during this hospitalization were excluded from analysis. Data collected included age, sex, symptoms and physical findings, laboratory and radiological investigations performed, co-morbidities, time from arrival to the hospital to operation, admission service, complications, and length of stay (LOS). Patients presenting with discharge diagnosis of acute cholecystitis but not undergoing cholecystectomy during the hospitalization were excluded.

In order to help account for the contributions of pre-operative comorbid illnesses to the measured outcomes, we classified co-morbidities as none, minor (abnormal electrocardiogram (ECG), rhythm other than sinus, low functional capacity, history of stroke, uncontrolled hypertension), intermediate (mild angina, previous myocardial infarction, compensated or prior heart failure, diabetes mellitus, renal insufficiency), or major (unstable coronary syndromes, decompensated heart failure, significant arrhythmias, severe valvular disease) following the criteria by American Heart Association/American College of Cardiology (AHA/ACC) for cardiovascular risk[11]. An abnormal ECG is one indicating left ventricular hypertrophy, left bundle branch block, ST-T abnormalities. Advanced age is also listed in the AHA/ACC criteria as a minor clinical predictor but is undefined. For this study, it was arbitrarily designated as > 75 years old. Liver cirrhosis, which is not included in the AHA/ACC criteria, was added as a major peri-operative risk factor because it is known to increase post-operative mortality[1213].

Post-operative complications were defined as newly arising occurrences related to the patient’s surgical disease or management that prolonged length of stay or required treatment. Complications were categorized as mild (e.g., urinary tract abnormalities, prolonged ileus, prolonged biliary drainage), moderate (e.g., systemic inflammatory response, deep wound infection), and severe (e.g., death, respiratory failure requiring intubation, myocardial infarction)[14].

A total of 88 patient charts with the discharge diagnosis of acute cholecystitis were reviewed. In six patients the pathology report did not confirm a diagnosis of acute cholecystitis, and it was concluded that they suffered from biliary colic. An additional 12 patients were discharged without cholecystectomy and were also felt to have had biliary colic rather than acute cholecystitis. These 18 patients were excluded from analysis, yielding a study population of 70 patients. The mean ± SD age was 53 ± 17.3 years. The majority of patients (69%) had no recognized peri-operative co-morbidities (Table 1). Among those with co-morbidities, 16% were classified as minor, 13% were intermediate, and 3% were major. All but one patient presented with abdominal pain and nearly all (89%) had right upper quadrant tenderness on examination. Murphy’s sign, considered to be a relatively specific sign in cholecystitis, was present in only one-quarter of the patients. Fever (> 37.8°C and leukocytosis), which together with Murphy’s sign comprise the diagnostic triad for acute cholecystitis, was present in only 23% and 51% of patients, respectively. Guarding was noted in 11% and rebound tenderness was detected in one patient.

Eleven patients presented with aspartate amino-transferase (AST) and alanine aminotransferase (ALT) greater than three times the upper normal limit and 2 (18%) required pre-operative ERCP for stone extraction. Two other patients had pre-operative ERCP performed because of common bile duct dilatation identified on CT scan. All ERCP examinations were therapeutic, removing stones from the common bile duct via sphincterotomy. Three additional patients required post-operative ERCP for retained stones.

The majority of patients were admitted to the surgical service (83%) (Tables 1 and 2). Patients admitted to the medical service were significantly older (66 years vs 52 years, P < 0.05), and had more minor and intermediate, but not severe, peri-operative cardiovascular risk factors (P < 0.05), including diabetes mellitus, prior heart failure, and age > 75 years. Other non-cardiovascular comorbid conditions among the patients admitted to the medical service included dementia, anxiety, brain tumor, hepatitis B, and chronic Clostridium difficile infection.

Patients admitted to the medical service had significantly more abdominal CT scans done in the evaluation process (Table 2). The distribution of other pre-operative tests was not significantly different between the surgical and medical services. There was a trend toward more medicine patients undergoing pre-operative ERCP, 2 of 12 (17%) patients vs 2 of 58 (3%) patients, though this did not reach statistical significance.

The mean time to operation from admission was 39.3 ± 43.1 h (Table 2). Operation within 72 h occurred in 87% of the cases and operation within 24 h occurred in 51% of the cases. Patients admitted to the surgical service compared to the medical service had a significantly decreased mean time to operation, and significantly more surgical patients were operated on within 3 d as well as within 24 h (Table 2).

Post-operative complications occurred in 16 cases (23%). Complications occurred more frequently in medicine patients: 7 of 12 (58%) versus 9 of 58 (16%) surgical patients, P < 0.01 (Table 2). Complications included prolonged (> 2 d) ileus (n = 2), discharge with Jackson-Pratt drain for biliary drainage (n = 2), urinary retention requiring discharge with foley catheter (n = 2), volume resuscitation for systemic inflammatory response leading to post-operative volume overload requiring diuresis (n = 1), penile edema (n = 1), gout (n = 1), pancreatitis (n = 2), delayed operation for gangrenous cholecystitis leading to post-operative deconditioning with inability to walk (n = 1), preoperative antibiotics to “cool down” cholecystitis associated with diffuse rash requiring steroids (n = 1), hypoxic respiratory failure (n = 1), cardiac arrest leading to death (n = 1), and multi-organ system failure leading to death (n = 1).

Mean ± SD LOS for all patients was 5.0 ± 3.9 d. Patients admitted to the surgical service experienced a shorter LOS compared to patients admitted to the medicine service (Table 2). LOS of 3 d or less was seen more often for surgical patients, while LOS of 6 d or more was observed most often for medicine patients.

Using hierarchical generalized linear models, we evaluated the impact of various factors on time to surgery (Table 3). Adjusting for patient demographics (age and gender), peri-operative co-morbidity level, physical signs, and pre-operative tests, we found that the performance of pre-operative ERCP was associated with increased time to operation. Select physical and laboratory signs (Murphy’s sign, temperature > 37.8°C, WBC > 10.2 thousand/UL) had no apparent impact on time to operation in both the unadjusted and adjusted models. Medical admitting team was significantly associated with increased time to operation until ERCP was included in the model (Table 3).

The number of pre-operative diagnostic imaging studies was significantly associated with increased time to operation (Table 3). Time to surgery doubled as the number of diagnostic procedures increased from one to 2 or 3 tests and doubled again when 4 tests were performed. Adjusted models to determine the association between time to surgery and individual studies and combinations of studies revealed that ultrasound alone was associated with the shortest time (mean 23.3, 95% CI 10.7-35.9 h) and the combination of CT scan and MRCP with longest time (mean 195.4, 95% CI 125.1-256.8 h) to surgery (Table 4). A trend is present toward decreased risk of complication associated with surgery within 24 h of hospital arrival that becomes non-significant after adjustment for diagnostic imaging studies (Table 5).

In this study, we find that 87% of cholecystectomies occurred within 72 h of arrival at the hospital and half of them within 24 h, indicating an overall timely approach. Several studies have shown that performing LC within 72 h of presentation does not increase morbidity and mortality[69]. Early LC, furthermore, has been shown to have a lower complication rate, decreased length of stay, decreased cost, and a shorter post-operative recovery period compared to delayed LC[5781617]. The results from a meta-analysis by Papi et al of 12 prospective randomized trials from 1970-2000 (9 addressing open cholecystectomy, 3 addressing LC), showed that there was a decrease in conversion to open procedure when LC was done early (17.6% vs 25.7%) and there were fewer post-operative complications (10.9% vs 15.6%)[17]. Another benefit of early surgery proved to be a shorter length of hospital stay (6.3 d vs 9.9 d).

Our analysis suggests a number of potential ways to speed up time to cholecystectomy. The number of pre-operative procedures was significantly associated with adjusted time to surgery, especially when four or more tests were done. Performance of MRCP and CT scan had by far the longest associated adjusted time to surgery. While the need for ERCP and MRCP will always result in some delay in surgery, these studies are routinely available within 24 h at our institution, often on the same day as requested. Efforts to streamline the decision process leading to ERCP and MRCP and then providing these studies promptly when requested may lead to earlier surgery for patients with acute cholecystitis. Additionally, more judicious use of pre-operative diagnostic imaging studies would also reduce delay. Although the sample size was small, not surprisingly, the performance of ERCP was also associated with delayed time to operation.

Also noted was that patients admitted to the medical service experienced over twice the wait for surgery compared to those going to the surgical service. No doubt some of this delay represents treatment of co-morbid illnesses, which were significantly more common in patients admitted to the medical service. However, patients on the medical service also had more imaging procedures performed, mostly abdominal CT scans, indicating confusion over the diagnosis and contributing to delays in definitive surgical therapy. Surgeons directly caring for patients with abdominal pain may be more apt to make the early diagnosis of acute cholecystitis with fewer imaging studies than medical physicians and, thereby, bring those patients to operation earlier. Although admitting patients with abdominal pain suspicious for cholecystitis to the surgical service may reduce time to cholecystectomy, it must be weighed against the advantages of having patients with active co-morbidities managed on the medical service with surgical consultation.

Our findings raise the question whether the benefit seen by early surgery is conferred mainly on those patients operated on within 24 h of arrival. A significant decreased risk of complications was seen in this subset of patients, except in the model including adjustment for diagnostic imaging studies. A reasonable inference is that this subset of patients may be so straight forward that little diagnostic uncertainty exists, fewer tests are required, and admission to the surgical service is favored. Conversely, those patients about whom diagnostic uncertainty is present or who are chosen for “cooling off” with pre-operative antibiotics may be placed at greater risk of complications because of the prolonged inflammatory state. Acute inflammation creating edema in the gallbladder submucosa can mature into a chronic inflammatory response with neovascularization and fibrosis, which can make a delayed procedure potentially more difficult[18]. Often, an operation will be deferred, especially in patients with pre-existing co-morbidities, in an attempt to maximize a patient’s medical tolerance of an operative insult. Firm evidence is lacking for whether this delayed approach is truly the best. Indeed, a recent meta-analysis comparing LC within 7 d versus longer than 6 wk after index admission showed no clear benefit from waiting[19]. Although the complications observed in the patients in this analysis can be directly or indirectly linked to acute cholecystitis and surgery, no causal connection between them and any delay in operating can be drawn. A prospective trial is likely required to determine the optimal times for LC.

Somewhat surprising was that what many would deem a classic clinical presentation with fever, leukocytosis, and Murphy’s sign was not associated with time to surgery. In a retrospective study of patients undergoing nuclear hepatobiliary scans by Singer et al[20], fever had a positive predictive value of 100% and a sensitivity of 14.6% for a positive scan. A positive Murphy’s sign had 97.2% sensitivity and 93.3% positive predictive value. In their systematic review of presenting signs and symptoms that may warrant investigation for cholecystitis, Trowbridge et al[21] found that no clinical or laboratory finding had a likelihood ratio (LR) sufficient to rule in or out cholecystitis. Murphy’s sign, however, appeared the most useful, with a positive LR of 2.8, 95% CI 0.8-8.6. In our study, possibly the variable skill of the clinicians assessing for cholecystitis and the ready availability of imaging studies mitigated any diagnostic benefit from the presence of these findings that would have sped up care.

We chose in this study to use the AHA/ACC risk classification scheme rather than the American Society of Anesthesiologists-Physical Status (ASA-PS) score for risk stratification. While the ASA-PS score’s simplicity and widespread use makes it an attractive index of risk, it is subjective and not explicitly intended for estimation of operative risk[22]. Furthermore, we had no measure of inter-anesthesiologist agreement on ASA-PS score and numerous anesthesiologists, so we could not control for variations at the level of the grader. While the AHA/ACC guidelines were derived for cardiovascular risk, it contains a detailed list of factors plausibly conferring risk for non-cardiac complications[23–25].

In summary, our analysis indicates good adherence to the desirable practice of performing cholecystectomies within 3 d of presentation to the hospital and raises the question of whether even prompter surgery may reduce the risk of complications. Efforts to expedite cholecystectomy should consider factors that may contribute to delay in surgery such as ERCP and MRCP, the combination of CT scan and MRCP, and admission to a medical service.