An audit of ambulatory laparoscopic cholecystectomy in a Singapore institution: are we ready for day-case laparoscopic cholecystectomy?

Abstract

Background. Day-Case laparoscopic cholecystectomy (LC) is practiced in many countries. However, this has yet to be widely accepted in Singapore. This study aims to determine the potential success rate of day-case LC in our institution. Patient and methods. We retrospectively assessed the proportion of our Ambulatory Surgery 23 hour (AS23) LC patients that met discharge criteria. Our proposed same-day discharge criteria include minimal pain, ability to tolerate feeds, ambulate independently and void spontaneously after 6–8 hours of monitoring. Results. From January 2005 to December 2006, of 405 patients listed for elective LC, 84% of patients were admitted to our AS23 ward. Patients with previous biliary sepsis or pancreatitis or who need laparoscopic common bile duct exploration (LCBDE) were included. The other 66 were admitted as inpatient. Forty-one of them were admitted due to conversion. A history of cholecystitis or cholangitis was a significant predictor of conversion to open surgery (OR=5.73 and 5.74 respectively, p<0.001). Of the 339 patients, 66% of them fulfilled all four criteria within eight hours of monitoring. Therefore, based on an intention-to-treat analysis, 51.2% fulfilled all four criteria and could potentially be discharged the same day. No predictor for failure was identified, including presence of co-morbidities, duration of operation, surgeon's grade and additional procedures like LCBDE. Conclusion. Using our current inclusion criteria, we projected a success rate of at least 50% with the implementation of day-case LC. With the attendant advantages of cost savings and reduced resource utilization, it is therefore worthwhile to start it in Singapore.

Introduction

Laparoscopic cholecystectomy (LC) has now become the standard treatment for symptomatic gallstone disease. Because of the smaller scars and reduced postoperative pain, patients enjoy a shorter hospital stay. Consequently, many healthcare providers have started to explore the feasibility of offering LC as a day-case procedure. This push towards day-case LC is also motivated in part by the drive to contain healthcare costs and relieve the strain on inpatient hospital beds. Indeed, many studies have documented the safety and feasibility of day-case LC in selected patients ¹,²,³,⁴,⁵,⁶,⁷,⁸. Four randomized trials comparing day-case versus overnight stay for LC have demonstrated that such a practice is evidence-based ⁹,¹⁰,¹¹,¹². From the economics point of view, studies have reported a cost savings of 11–46% for day-case LC as compared with inpatient treatment ⁵,⁷,⁸,¹³,¹⁴,¹⁵.

Ambulatory LC has now become routine practice in the USA ⁷ as well as many centers in Europe. In Singapore, however, there is still no published data of any local hospitals offering day-case LC despite the results achieved elsewhere. This is understandable as the implementation process of new services locally follows a cautious and step-wise manner. When LC was first introduced in our institution in 1991, patients then were admitted one-day prior to their operation, and stayed for one to two days postoperatively. With improvements in surgical and anaesthetic technique, the concept of same-day admission (SDA) was introduced in 2001, thereby shortening the length of stay (LOS) by one day. In 2005, our institution initiated the Ambulatory Surgery 23 hours (AS23) concept for patients undergoing certain elective operations of which LC was one of them. Under AS23, patients come on the day of their operation, and are discharged the following morning after a night of monitoring in the day surgery (DS) ward. They stayed in the hospital for less than 24 hours.

It is timely to review our current practice of AS23 service and evaluate our readiness for the next level of ambulatory care in our institution. The aim of this study is to review our patients who had undergone LC under AS23 and to assess the proportion of patients who fulfilled the discharge criteria for day-case LC. The data would also help us to project the potential success rate for day-case LC if we were to implement this scheme in our institution. Our secondary end-point is to identify factors that predict failure of same-day discharge.

Patient and methods

We reviewed the hospital records of all patients that were listed for an elective LC under AS23 in our hospital, a 1200-bedded tertiary teaching hospital, between January 2005 and December 2006. Based on the selection criteria of AS23, patients with an American Society of Anaesthesiologists (ASA) score of I or II could be listed for LC under AS23. Selected ASA III patients whose medical co-morbidity was deemed to be well-controlled were also included. Those with a body mass index (BMI) of more than 35 were excluded. Age, the clinical suspicion of common bile duct stones, a history of biliary sepsis, either cholecystitis or cholangitis, and a history of biliary pancreatitis were not the exclusion criteria for AS23 LC. The patients must have a responsible adult to accompany them home the following morning. As Singapore is only 42 km in her longest length, a prerequisite of patients having to live within a 50 km radius from the hospital as recommended by some authors is not relevant to Singapore ⁷,¹². All emergency cholecystectomies were excluded from this study.

Presence of gallstones was confirmed preoperatively using ultrasonography and on occasion, computer tomography was performed. We routinely performed liver function test (LFT) together with radiological imaging to predict the presence of bile duct stone. For patients with biliary ductal stones, they were offered the options of either a two-stage procedure in the form of endoscopic removal of stone followed by LC, or a single-stage procedure of LC with laparoscopic common bile duct exploration (LCBDE).

General anaesthesia was induced with propofol at 2–3 mg/kg and fentanyl at 1–2 mcg/kg. Muscle relaxation was achieved with atracurium 0.5 mg/kg and mechanical ventilation performed via endotracheal tube. Anaesthesia was maintained with sevoflurane in an air and oxygen mixture. At the end of the operation, atropine and neostigmine were used for reversal. Prophylaxis against postoperative nausea and vomiting (PONV) was achieved with intravenous droperidol at 20 mcg/kg at induction and a dose of intravenous ondansetrone 4 mg at reversal. Preemptive analgesia was not a routine practice in our institution, but rather, according to the preference of the individual anaesthetist. All port sites were infiltrated with 0.5% bupivacaine local anaesthetic at the end of the operation.

All patients underwent LC using a standard four-port technique (three 5 mm ports and one 10 mm port) with intra-abdominal pneumoperitoneum pressure set at 12 mmHg. We practiced selective intra-operative cholangiogram (IOC) using a transcystic technique. Our indications of IOC include any patients with history of jaundice, biliary pancreatitis or cholangitis with no prior clearance of biliary tree. Dilated biliary tree on any imaging modality and deranged LFT on most recent biochemistry (Alkaline phosphatase >3X normal limit or hyperbilirubinaemia) would also require IOC for bile duct clearance.

When LCBDE was indicated, this was performed via a transcystic route using the Nathanson Common Bile Duct Exploration Set (CDES-550-Nathanson, Cook Australia). Should the transcystic method fail, laparoscopic choledochotomy would be performed provided the Common Bile Duct (CBD) diameter was at least 1 cm. All patients received a dose of Cefazolin 1 g at induction as part of our Surgical Site Infection (SSI) prevention protocol. Closed suction drain may be placed at Morrison's pouch through the right lateral 5 mm port site at the discretion of the surgeon.

After initial recovery in our Post Anaesthesia Care Unit (PACU), they were transferred to our DS unit for further monitoring. Postoperative analgesia was provided using a standard package of oral paracetamol and a non-steroidal anti-inflammatory drug (NSAID) of the surgeon's choice, together with parenteral pethidine for breakthrough pain. Nausea or vomiting was treated with metoclopramide. Pain was assessed using a 10-point visual analogue scale (VAS) (0, no pain at all; 10, unbearable pain). This was done at intervals of 30 minutes for the first hour upon arrival at the DS ward, hourly for the subsequent five hours, and thereafter, four hourly until the point of discharge. In addition, the time when patient first took their oral feeds, first passed urine spontaneously and first ambulate independently were recorded.

For the purpose of this study, we defined the following criteria as prerequisites for same-day discharge after 6–8 hours of monitoring in the DS ward: (1) pain score of three or less on the VAS and requiring only oral analgesia; (2) ability to tolerate oral feeds; (3) ability to pass urine spontaneously; and (4) ability to ambulate independently.

Data were analyzed with Stata version 9.0 (Stata Corp, Texas, USA). All Statistical tests were carried out at 5% level of significance. Multiple logistic regression was applied to identify predictive factors.

Results

Between January 2005 and December 2006, a total of 405 patients were scheduled for elective LC under AS23 in our institution. Patient demographics and co-morbidities are shown in Table I. Majority (85%) of our patients were Chinese. This may represent the ethnic mix of our local population, where Chinese is the largest racial group, rather than a racial predisposition to gallstone disease. Fifty-six percent (55.6%) of patients had associated co-morbidity. All our patients in this study were ASA class I and II, except for three patients with ASA III. They had a history of previous myocardial infarction more than six months prior to the surgery but recovered uneventfully. We have included psychiatric disorders in our study as this may have an impact on patient's anxiety and hence perception of well-being postoperatively.

Table I. Patient demographics and co-morbidities (n=405).

Median age (years)	51	Range 15–92
	n	%
Gender
Male	175	43.2
Female	230	56.8
Race
Chinese	345	85.2
Malay	40	9.9
Indian	13	3.2
Others	7	1.7
Co-morbidities
Ischemic heart disease	19	4.7
Respiratory disorders	25	6.2
Diabetes mellitus	51	12.6
Hypertension	121	29.9
Renal disease	5	1.2
Psychiatric disease	8	2.0

The most frequent indication for LC was biliary colic (62.7%) (Table II). There were four patients who had gallstones detected incidentally and who requested for LC. Interval LC was performed for 96 patients who had previous acute cholecystitis and 26 patients who had previous cholangitis. Our protocol for interval cholecystectomy is to time the LC six weeks after the acute cholecystitis for those patients in whom emergency LC could not be performed for some reason or other. For cholangitis, the LC was timed within two weeks of biliary decompression. When we grouped those who had acute cholecystitis and cholangitis together, about one-third of patients had a previous history of biliary sepsis (30.1%).

Table II. Indications for LC and the nature of procedures (n=405).

	n (%)
Indications
Biliary colic	254 (62.7)
Previous acute cholecystitis	96 (23.7)
Previous cholangitis	26 (6.4)
Previous biliary pancreatitis	20 (4.9)
Incidental gallstone	4 (1)
Gallbladder polyp	5 (1.2)
History of biliary sepsis (cholecystitis and/or cholangitis)
Yes	122 (30.1)
No	283 (69.9)
Nature of procedure
LC	326 (80.4)
LC with IOC	33 (8.2)
LC with transcystic CBDE	5 (1.2)
Conversions to open operation	41 (10.1)

Note: IOC, intra-operative cholangiography; CBDE, common bile duct exploration.

Sixty percent of operations were performed by Consultants while the remainder was performed by advanced surgical trainees (ASTs). The mean operating time was 91±49 minutes. Our conversion rate was 10.12%, and the reasons are shown in Table III. All the patients who had conversion due to dense adhesions in the Morrison's pouch had prior inflammatory process such as cholecystitis and cholangitis. There was one case of common hepatic duct injury, and a hepatico-jejunostomy had to be constructed. This occurred in a patient with a previous history of cholangitis, and dense adhesions were encountered in the Calot's triangle. There was one postoperative mortality. This patient had Mirrizzi's syndrome, and had to be converted to open cholecystectomy and CBDE. Unfortunately, she developed severe necrotizing pancreatitis postoperatively and died subsequently from multi-organ failure.

Table III. Reasons for conversion (n=41).

	n
Dense adhesions	35
Bleeding	2
Bile duct injury	1
Bile leak from cystic duct stump	1
Inability to secure a large cystic duct	1
Cholecysto-colonic fistula	1

We attempted to identify risk factors that predict conversion. Using logistic regression analysis, the history of previous acute cholecystitis (OR=5.73, p<0.001, confidence interval, 2.92–11.22) and acute cholangitis (OR=5.74, p<0.001, confidence interval, 2.37–13.92) was significantly associated with conversion to open surgery. All other factors including age, co-morbidities, duration of surgery and surgeon's grade were not significant.

The outcome of the 405 patients listed for LC under AS23 is shown in Table IV. There were 66 (16.3%) patients who were converted from AS23 to SDA. Forty one of them (62%) were because of conversion to open surgery. For the other 25 patients, despite the successful removal of the gallbladder laparoscopically, they were converted to inpatient for various reasons (Table IV). Consequently, 339 patients were admitted into our DS ward. Of these, 97.9% were discharged the following morning as scheduled. The reasons for admission for the seven patients were patient's request (n=2), pain (n=3), high drain output (n=1) and hypotension (n=1). Drains were placed in 22 of these 339 patients. The median (range) drain output was 35 mls (0–120 mls). Eighteen of them had their drains removed the next morning prior to discharge, whilst three went home with the drain, after being counseled on drain care. An early review date was given to these three patients. As for the rest, a visit to our specialist clinic was scheduled six weeks after the operation.

Table IV. Outcome of 405 patients listed for elective LC under AS23.

	n (%)
Outcome variables
Listed for elective LC under AS23	405 (100)
Conversion to SDA (inpatient admission)	66 (16.3)
Admitted to DS ward under AS23	339 (83.7)
Reasons for conversion to SDA (n=66)
Conversion to open surgery	41 (62.1)
Surgeon's preference	7 (10.6)
Patient's request	6 (9.1)
Medical reasons:	12 (18.2)
Medical observation because of co-morbidity	11
Chest pain in PACU	1
Outcome of Patients admitted to AS23 (n=339)
Discharged the following morning	332 (97.9)
Converted to inpatient the following morning for further monitoring	7 (2.1)

Note: SDA, same-day admission; DS, day surgery.

In projecting the potential success rate of day-case LC, we applied our discharge criteria for same-day discharge to the 339 patients who were admitted to the DS ward as scheduled under AS23 (Table V). More than 90% of our study population have a low pain score (≤3) and were able to tolerate oral feeding without PONV within eight hours of monitoring in the DS ward. When all four criteria were applied, 221 patients fulfilled the discharge prerequisite for same-day discharge. Based on an intention-to-treat (ITT) analysis, potentially 52.1% of our original 405 patients would be eligible for same-day discharge. By univariate analysis of the factors between those who fulfilled all four discharge criteria and those who did not, we were not able to identify any risk factors that predict failure to meet our same-day discharge criteria. This includes patient's age, gender, race, co-morbidities, duration of operation, surgeon's grade, the need for IOC or CBDE, the presence of drains, and a history of acute cholecystitis or cholangitis. There was no unplanned readmission in our study population.

Table V. Proportion of patients meeting the pre-set discharge criteria for same-day discharge (n=339).

	n (%)
Number of patients with a pain score on the VAS less than or equal three after eight hours of observation	306 (91.6)
Number of patients who could tolerate oral feeds after eight hours of observation	310 (91.7)
Number of patients who could pass urine spontaneously after eight hours of observation	272 (81)
Number of patients who could ambulate independently after eight hours of observation	252 (74.9)
Number of patients who fulfilled all four discharge criteria after eight hours of observation	221 (63.4)
Potential success rate of day-case LC in our institution (based on ITT analysis)	221/405 (52.1)

Note: VAS, visual analogue scale; ITT, intention-to-treat.

Discussion

The present study of our AS23 LC was undertaken to assess the feasibility of offering day-case LC in our institution. Based on an ITT analysis, without changing our care processes, including patient selection, we should expect a success rate of at least 51.2% should we offer day-case LC. This seemed low when compared to other published figures of about 80% ²,⁵,⁷. Patient selection has a major impact on the success rate of a day-case LC program ¹⁶,¹⁷,¹⁸. Hence, many centers would exclude patients who have a history of cholecystitis, cholangitis or biliary pancreatitis from day-case LC ⁵,⁸,¹². Such bias is supported by a study which found that a previous diagnosis of acute cholecystitis or biliary pancreatitis was highly predictive of hospital admission ¹⁹. A clinical suspicion of common bile duct stone was also a contraindication for day-case LC ³,¹²,²⁰.

Under our current AS23 inclusion criteria, we have included patients with a previous history of cholecystitis, cholangitis and biliary pancreatitis. This is mainly because our current system allows an overnight observation for such patients. However, from this study, whilst a previous history of biliary sepsis (cholecystitis or cholangitis) is a significant predictor of conversion to open surgery, those who managed to have their gallbladder removed laparoscopically were as likely to fulfill all discharge criteria for day-case LC as those with no previous history of biliary sepsis. Furthermore, we did not have unplanned readmission. In any case, most postoperative complications of the procedure occur more than 24 hours after surgery ²¹,²². Hence, an overnight stay does not necessarily improve safety and may be avoided in motivated, fit patients ¹,²³. In patients with suspected or confirmed choledocholithiasis, the operation is expected to be more difficult and hence longer in duration. The duration of operation was found by various authors to be predictive of successful early postoperative discharge ¹⁸,²⁴. However, in the series by Narain et al. they found otherwise ¹. Similarly, in our series, even the addition of cholangiography or additional procedures did not increase the likelihood of the patient staying overnight, although we acknowledge that the number of patients falling into this category is small (LC with IOC, n=33; LC with CBDE, n=5). Consequently, because of our less stringent selection criteria, our mean operating time is longer than the usual published figure for day-case LC ³,⁶,⁷. We did not observe a difference in operating times between that of a consultant and that of a trainee. This may not be a true reflection. This is because when a trainee encounters difficulties and calls in the consultant, the case in question will be logged to the consultant.

Despite having less stringent selection criteria, can our success rate be further improved for day-case LC? One way is to implement preemptive analgesia as part of our protocol. Preemptive analgesia with NSAIDs has an opiate-sparing effect, thereby potentially reducing the incidence of PONV ²⁵,²⁶,²⁷. In addition, antinociceptive treatment started before surgery is more effective in reducing postoperative pain than treatment started in the early postoperative period ²⁸. A second strategy is to offer the patients oral feeds, and encourage them to mobilize once they are admitted to the DS ward, modelling the practice in some centers ³,¹². Our current practice is to allow patients oral fluids on demand, and out of their beds on request. The time when patient can tolerate oral feeds, can ambulate independently and can void spontaneously is thus recorded as such. We believe a proactive approach can increase the proportion of our patients who can achieve these criteria by 6–8 hours of monitoring, up from the present figures of 91.7%, 74.95% and 81%, respectively in this study, hence potentially increasing the numbers of patients suitable for same-day discharge.

Finally, many authors have documented that a high patient satisfaction rate for day-case LC, ranging from 60 to 95% ⁹,²⁹,³⁰,³¹,³²,³³. However, in a multi-racial and multi-cultural country like Singapore, an accurate estimation of patient acceptance can only be obtained from a prospective quality-of-life analysis. Nonetheless, an essential ingredient for patient satisfaction is adequate preoperative and postoperative patient education. A higher level of anxiety can be expected in patients offered day-case LC, owing to the fear of suffering complications and pain at home ¹². Thus, it is imperative that should we embark on day-case LC, adequate information must be given to the patients on discharge. Measures used by some centers include information leaflets providing advice on the expected recovery pattern, pain control, recognition of complications and return to activities ³,⁴. Patients should also be contacted within 24 hours of discharge to ascertain their adequate recovery. Finally, patients should be given the telephone numbers of our nursing staff and a protocol for expedited hospital admission when the need arises, rather than through a busy emergency department, must be available.

Conclusion

In conclusion, based on this study, even without making any changes to our current inclusion criteria, we should expect a success rate of at least 50% should we embark on day-case LC. This paper prompts us to give day-case LC a serious consideration. Given the escalating healthcare costs and increasing capacity constraints in many countries, day-case LC is a good strategy in keeping inflationary pressures down. Singapore is ready to embark on day-case LC.