Objective The optimal management of acute cholecystitis in patients at very high risk for cholecystectomy is uncertain. The aim of the current study was to compare endoscopic ultrasound (EUS)-guided gallbladder drainage (EUS-GBD) to percutaneous cholecystostomy (PT-GBD) as a definitive treatment in these patients under a randomised controlled trial.
Design Consecutive patients suffering from acute calculous cholecystitis but were at very high-risk for cholecystectomy were recruited. The primary outcome was the 1-year adverse events rate. Secondary outcomes include technical and clinical success, 30-day adverse events, pain scores, unplanned readmissions, re-interventions and mortalities.
Results Between August 2014 to February 2018, 80 patients were recruited. EUS-GBD significantly reduced 1 year adverse events (10 (25.6%) vs 31 (77.5%), p<0.001), 30-day adverse events (5 (12.8%) vs 19 (47.5%), p=0.010), re-interventions after 30 days (1/39 (2.6%) vs 12/40 (30%), p=0.001), number of unplanned readmissions (6/39 (15.4%) vs 20/40 (50%), p=0.002) and recurrent cholecystitis (1/39 (2.6%) vs 8/40 (20%), p=0.029). Postprocedural pain scores and analgesic requirements were also less (p=0.034). The technical success (97.4% vs 100%, p=0.494), clinical success (92.3% vs 92.5%, p=1) and 30-day mortality (7.7% vs 10%, p=1) were statistically similar. The predictor to recurrent acute cholecystitis was the performance of PT-GBD (OR (95% CI)=5.63 (1.20–53.90), p=0.027).
Conclusion EUS-GBD improved outcomes as compared to PT-GBD in those patients that not candidates for cholecystectomy. EUS-GBD should be the procedure of choice provided that the expertise is available after a multi-disciplinary meeting. Further studies are required to determine the long-term efficacy.
Trial registration number NCT02212717
- EUS-guided gallbladder drainage
- percutaneous cholecystostomy
- acute cholecystitis
- interventional endoscopic ultrasonography
- lumen apposing metal stents
Statistics from Altmetric.com
- EUS-guided gallbladder drainage
- percutaneous cholecystostomy
- acute cholecystitis
- interventional endoscopic ultrasonography
- lumen apposing metal stents
What is already known on this subject?
EUS-guided gallbladder drainage (EUS-GBD) has gained attention as a potential method of internally draining the gallbladder and removing gallstones in high-risk patients. Several retrospective studies and meta-analyses have shown that EUS-GBD may be associated with improved outcomes over percutaneous cholecystostomy.
What are the new findings?
In this randomised trial of 80 patients, EUS-GBD significantly reduced 1-year adverse events (25.6% vs 77.5%, p<0.001), 30-day adverse events (12.8% vs 47.5%, p=0.001), re-interventions after 30 days (2.6% vs 30%, p=0.001), number of unplanned readmissions (6% vs 50%, p=0.002) and recurrent cholecystitis (2.6% vs 20%, p=0.029).
How might it impact on clinical practice in the foreseeable future?
EUS-GBD improved outcomes as compared to PT-GBD in those at very high risk of cholecystectomy. This should be the procedure of choice in experienced centres.
The incidence of acute cholecystitis in patients with gallstones increases with age. The gold standard of treatment is laparoscopic cholecystectomy (LC).1 2 If the patient is not suitable for early surgery, then percutaneous cholecystostomy (PT-GBD) should be performed followed by delayed LC once the patient’s overall condition has improved. However, many elderly patients suffer from multiple co-morbidities that render them not candidates for cholecystectomy. Current guidelines do not provide definite recommendations on how to manage patients that are at very high risk for cholecystectomy and would never be considered for surgery. Patients would have their gallbladders drained percutaneously if sepsis cannot be controlled and the optimal management after drainage without the option of cholecystectomy remains undefined.1 3
Recently, EUS-guided gallbladder drainage (EUS-GBD) has gained attention as a potential method of internally draining the gallbladder and removing gallstones in high-risk patients.4–12 Several comparative studies have suggested that EUS-GBD may be associated with improved outcomes compared to PT-GBD.4–6 However, results from randomised studies with predefined protocol and outcome parameters are lacking. Hence, the aim of the current study is to compare the outcomes of EUS-GBD versus PT-GBD in treatment of acute calculous cholecystitis in patients that are at very high-risk for cholecystectomy under a randomised setting, to provide an unbiased assessment on the efficacy of both interventions in the study population. We hypothesise that EUS-GBD would reduce 1 year adverse events compared to PT-GBD.
This was a prospective multi-centred open labelled randomised controlled study involving five high-volume institutions. Diagnosis and severity of acute calculous cholecystitis were defined according to the Tokyo guidelines 18 (online supplementary table 1).13 Consecutive patients aged ≥18 years old and suffering from grade 2 and 3 acute calculous cholecystitis but were at very high-risk for cholecystectomy due to poor premorbid conditions were recruited (online supplementary table 2). Patients were deemed very high-risk for cholecystectomy if they satisfied one or more of the following criteria: Age ≥80 years old, American Society of Anesthesiology grade 3 or above, age-adjusted Charlson Comorbidity Index >5 and/or Karnofsky score <50. The age-adjusted Charlson Comorbidity Index is composite score that takes into account of the comorbidity conditions of the patient that predicts the 1-year mortality risk of the patients.14 The surgeons and the anesthesiologist jointly made the decision in the respective institutions regarding which patients were deemed to be at very high-risk for cholecystectomy. Elderly patients suffering from acute cholecystitis but refused operations were also included. The inclusion and exclusion criteria are shown in online supplementary table 2. Included patients were randomly allocated to receive EUS-GBD or PT-GBD within 4 to 6 hours after diagnosis. None of the patients subsequently received cholecystectomy. The study was conducted according to the Declaration of Helsinki and the International Conference on Harmonisation, good clinical practice guidelines. All authors had access to the study data and reviewed and approved the final manuscript.
The patients were not involved in design of the study as many of them are old and frail and could not understand or participate in the process.
Prior to the procedures, the patients were kept fasted for 6 hours and given intravenous antibiotics (second generation cephalosporin or equivalent). After the study intervention, antibiotics were continued for up to 1 week after the procedure or longer depending on the clinical course of the patient.
EUS-guided gallbladder drainage
All procedures were performed by an experienced endosonographer who has performed more than 25 EUS-GBD (the number of procedures suggested to be required to gain competency in one paper).15 The gallbladder was identified by a linear echoendoscope (GF-UTC 260 Olympus, Tokyo, Japan) and a suitable puncture site in the stomach or the duodenum without intervening blood vessels was located. EUS-GBD was performed by either the conventional or direct method (figure 1). In the conventional method, the gallbladder was punctured with a 19-gauge needle and the position was confirmed by aspiration of bile and/or minimal contrast injection. A 0.025-inch or 0.035-inch guidewire was passed through the needle and looped in the gallbladder. The delivery system of the cautery-enhanced lumen apposing metal stent (LAMS) (AXIOS, Boston Scientific Medical Corporation, Marlborough, USA) was inserted. In the direct method, the delivery system of the stent was directly inserted into the gallbladder without prior needle insertion. A 10×10 mm stent system was used if the largest gallstone was smaller than 10 mm in size and a 15×10 mm stent was used if the largest gallstone was larger than 10 mm. The distal flange of the stent was deployed under EUS guidance, followed by deployment of the proximal flange under endoscopic guidance. Once deployed, the gallbladder was completely emptied by suction and irrigation until the effluent through the stent is clean. The decision to insert an additional double pigtail plastic stent through the LAMS was at the discretion of the endosonographer.
Experienced interventional radiologists who have performed more than 50 PT-GBD in the respective hospitals performed the procedure under local anaesthesia.16 The transhepatic route was preferred in all patients to decrease the risk bile leakage. Otherwise, the gallbladder was drained by direct transperitoneal drainage. An 8.5 Fr pigtail drainage catheter (Cook Medical, USA) was placed between the eighth or ninth intercostal space under sonographic and fluoroscopic guidance. The pigtail catheter was drained to a bedside bag.
Failure of study intervention
If drainage of the gallbladder could not be achieved with the randomised study intervention, the reasons leading to such were recorded. The patients were then allowed to crossover to the alternative treatment arm for an attempt to perform GBD.
Patients were monitored after the procedure for signs of improving or worsening sepsis. If resolution of signs and symptoms were not reached after the 96 hours, then patency of the tubes or stents were checked with bedside aspiration or endoscopy. If sepsis still failed to resolve, CT of the abdomen was performed to look for potential complications resulting from acute cholecystitis or the procedure. All patients were given oral acetaminophen tablets 500 mg and intravenous tramadol hydrochloride (0.7 mg/kg every 4 hours) or equivalent as required. Oral diets were resumed when the patient was afebrile and had presence of flatus or bowel output. The patients were discharged if they were afebrile and had >20% drop in white blood cells counts.
Patients that received EUS-GBD were scheduled for a follow-up per oral cholecystoscopy 1 month after the procedure to check for clearance of stones.12 The procedure was performed under conscious sedation. A regular endoscope was inserted through the gallbladder stent into the gallbladder to check for presence of gallstones. If all gallstones were cleared, the LAMS were removed with rat tooth forceps (FG-42L-1, Olympus, Tokyo, Japan) and replaced with a permanent 7 Fr double pigtail plastic catheter. If stones cannot be completely cleared, then patients are scheduled for another follow-up cholecystoscopy 1 month afterwards. This would be repeated until all stones were removed and then the LAMS were removed. In patients that were too old or frail to undergo additional endoscopic procedures, then the gallbladder stents were placed permanently.
Patients in the PT-GBD group received a tube cholecystogram 1 month after the procedure to check for the patency of the cystic duct. If the duct was patent, then the cholecystostomy were capped and left in-situ or removed. If the cystic duct were obstructed, then long-term cholecystostomy drainage was provided.
All patients were followed-up at 3, 6, 9 and 12 months after the procedure by the investigators. They were followed-up until 1 year or death. Any adverse events, re-interventions and unplanned admissions related to stone disease or the study intervention during the study period were recorded.
The primary outcome measurement was the cumulative rate of adverse events in 1 year. Secondary outcomes include technical and clinical success, daily postprocedural pain scores for the first 7 days, analgesic requirements, unplanned readmissions and reinterventions. Technical success was defined as the ability to access and drain the gallbladder by placement of a stent or drainage tube and maintenance of good drainage. Clinical success was defined as improvement in clinical symptoms and laboratory test. Pain assessment was performed using the visual-analogue scale on postprocedural days 1 to 7. Only biliary tract related adverse events after the 30-day period were recorded as many of these patients suffered from other co-morbidities that may require frequent admission. The severity of adverse events was graded according to Clavien-Dindo classification.17 The number of unplanned admissions due to gallstone related disease or the study interventions were recorded up to 1 year. The definitions of the outcome measurements are shown in online supplementary table 3.
The patients were randomised after inclusion into the study. The patients or investigators could not be blinded due to the nature of the intervention. The randomisation codes were stratified according to the participating hospital and generated in blocks of 10. The randomisation codes were obtained through a web-based Internet application by the participating site.
Sample size calculation
The sample size was calculated based on the mean percentage of 1-year adverse events from our prior study and other studies.5 6 Assuming a 31.7% difference in adverse events with a two-sided p value of 0.05 and a power of 80%, 35 patients were required in each group. Considering a 15% dropout rate, 40 patients would be required in each group.
All outcomes were analysed according to the intention-to-treat principle. Statistical analyses were performed using SPSS V.20.0 statistical software (SPSS, Chicago, Illinois, USA) and R software V.3.6.0. Comparisons were made by χ2 test or Fisher’s exact test for categorical data, and Mann-Whitney U test and t-test for continuous data, where appropriate. The predictors to 30-day adverse events rate were assessed using Firth logistic regression analysis. The Firth method provides bias-reduction for small sample size as well as yields finite and consistent estimates even in case of separation.18
From August 2014 to February 2018, 80 patients satisfying the inclusion criteria were included. One patient subsequently diagnosed of gallbladder cancer was excluded from the analysis. The consort flow chart is shown in figure 2. There were no differences in background demographics (table 1). The mean (SD) age-adjusted Charlson Comorbidity Index was 5.6 (1.6) in the EUS-GBD group and 5.6 (1.4) in the PT-GBD group. Grade 2 cholecystitis were present in 34 patients in both groups and grade 3 cholecystitis in five and six patients, respectively (EUS-GBD and PT-GBD). The median (IQR) duration of time from symptom presentation to procedure was 3 (2–4) days in the EUS-GBD group and 3 (2–4.75) days in the PT-GBD group (p=0.745). There were no statistically significant differences in the numbers of procedures that were performed in same day (EUS-GBD vs PT-GBD, 18/40 (45%) vs 14/39 (35.9%), p=0.495) or the next day. Thirty EUS procedures were performed under conscious sedation and nine under monitored anaesthesia. While all PT-GBDs were performed under local anaesthesia. None of the recruited patients ultimately received cholecystectomy.
The clinical outcomes of the two groups are shown in table 2. When comparing EUS-GBD to PT-GBD, there were no differences in technical success (38/39 (97.4%) vs 40/40 (100%), p=0.494) and clinical success (36/39 (92.3%) vs 37/39 (92.5%), p=1). One patient in the EUS-GBD group did not achieve technical success as there was no window for puncture from the stomach or duodenum. The patient was then subjected to PT-GBD but this also failed. Three patients in the EUS-GBD group and the PT-GBD group had clinical failures and all of these patients died. The 30-day mortality (3/39 (7.7%) vs 4/40 (10%), p=0.675) were not significantly different between the groups. In addition, EUS-GBD reduced postprocedural pain scores at the postprocedural day 0 (p=0.042), days 1 (p=0.013) and 2 (p=0.023) (figure 3). The total mean (SD) analgesic requirements were also less in the EUS-GBD group (3345 (5663) vs 5165 (5068), p=0.034). The mean (SD) procedural time (22.7 (13.0) vs 27.4 (12.0), p=0.108) and median (range) hospital stay (8 (4–13) vs 9 (7 – 14), p=0.181), were not significantly different statistically.
Furthermore, EUS-GBD significantly reduced 1-year adverse events (10 (25.6%) vs 31 (77.5%), p<0.001) and 30-day adverse events (5 (12.8%) vs 19 (47.5%), p=0.001). Eight patients in the EUS-GBD group and 12 patients in the PT-GBD group had ≥grade 3 adverse events after 1 year. The types of adverse events are shown in table 3. The majority of the 30-day and 1-year adverse events in the PT-GBD group were due to tube dislodgements. Significantly less patients in the EUS-GBD group had recurrent acute cholecystitis in 1 year (1/39 (2.6%) vs 8/40 (20%), p=0.029). 7.7% in the EUS-GBD group and 2.5% in the PT-GBD group had bile duct stones requiring endoscopic retrograde cholangiopancreatography (ERCP). One patient in the PT-GBD group suffered from recurrent cholecystitis and bile ducts stones in two episodes during the course of 1 year. Thus, the EUS-GBD group also had lower re-interventions for recurrent acute cholecystitis after 30 days (1/39 (2.6%) vs 8/40 (30%), p=0.001) and lower number of unplanned admissions (6/39 (15.4%) vs 20/40 (50%), p=0.002). The total recurrent biliary complication rate at 1 year was 10.3% in the EUS-GBD group and 20% in the PT-GBD group (p=0.227). The total number of scheduled and unscheduled re-interventions in each group were 33 (EUS-GBD) and 43 (PT-GBD). No difference in 30-day adverse events, 30-day mortality and 1 year adverse events were present in either group when analysing those patients that had procedures done on the same day vs the next day.
In the EUS-GBD group, 31 procedures (79.5%) were drained from the duodenum and 17 procedures (43.6%) were performed with the direct method. Thirty five EUS-GBD (89.7%) were performed with the 10×15 mm LAMS and additional plastic stents were placed through the LAMS in five patients. A follow-up per oral cholecystoscopy was performed 1 month after the procedure in 27 patients. The other patients refused repeat endoscopy as they were too old and frail and the stents were left permanently. Gallstones were spontaneously passed out from the gallbladder through the stent in 24 patients. Three patients required one additional procedure for endoscopic removal of the gallstones. In all patients where the gallstones were cleared, the LAMS were removed and replaced with a 7 Fr double pigtail plastic stents.
In the PT-GBD group, 15 patients (37.5%) were drained via the transhepatic route, and there was no window for transhepatic puncture in the others. Thirty six procedures (90%) were performed with a single puncture. Follow-up cholecystogram was performed in 27 patients (67.5%). The procedure was not performed in the remaining patients as the tube was already dislodged before the scheduled appointment. The cystic ducts were patent in 21 patients, 12 of 21 patients had their tubes plugged and the remaining had the tubes removed. Among those with tubes removed, 2 out of 9 patients developed recurrent acute cholecystitis.
The predictors to recurrent acute cholecystitis were then analysed by Firth logistic regression (table 4). The performance of PT-GBD (OR (95% CI)=5.63 (1.20–53.90), p=0.027) was a significant predictor to recurrent acute cholecystitis. A power calculation using a 5% level of significance, with performance of PT-GBD as the main predictor and adjusting for the effects of other covariates yielded the estimated power as 53.4%.
In the current study, we demonstrated that EUS-GBD and PT-GBD were both safe and effective methods of draining the gallbladder in patients suffering from acute cholecystitis. After 1 year, EUS-GBD reduced adverse events, recurrent acute cholecystitis, re-interventions and unplanned admissions. These findings support the use of this modality as a definitive treatment for acute cholecystitis in those patients that cannot receive cholecystectomy in experienced centres. Furthermore, the modality was also associated with short-term benefits of reduced 30-day adverse events, postprocedural pain scores and analgesic requirements.
The gold standard of treatment for acute cholecystitis is early LC or delayed LC after urgent percutaneous drainage.1 2 In the CHOLCOLATE trial, randomising patients with high-APACHE II scores to urgent LC or PT-GBD, cholecystectomy was associated with reduced complications (12% vs 65%, p<0.001), hospital stay (5 days vs 9 days, p<0.001) and reinterventions (12% vs 66%, p<0.001).16 This study suggests that even in critically ill patients, if they can withstand cholecystectomy, surgery is still preferred over PT-GBD. Cholecystectomy can also reduce the rates of recurrent cholecystitis in the future. However, the management strategy in those patients that could not tolerate cholecystectomy due to poor premorbid conditions is undefined. A common strategy would be to treat these patients conservatively and perform percutaneous drainage if sepsis cannot be controlled. However, management of the cholecystostomy tube is often cumbersome in these frail patients and the cholecystostomy is associated with a risk of adverse events between 16.2% and 25%.19–21 Previous attempts to remove gallstones after percutaneous cholecystostomy have also fallen out of favour, as the procedure requires repeated track dilations and associated with risk of recurrent stones even if stones were cleared.22 23
EUS-GBD is gaining popularity as an option of achieving internal drainage of the gallbladder in those that are at very high-risk for cholecystectomy.4–11 The results from a number of retrospective comparative studies have favoured EUS-GBD. Four studies have compared EUS-GBD to PT-GBD in patients with acute calculous cholecystitis.4–7 All reported comparable technical and clinical success rates between the two procedures. Two studies reported lower readmission rates, three studies reported lower re-intervention rates and one study reported significantly lower 1 year adverse events rates. A meta-analysis then concluded that EUS-GBD also had fewer adverse events than PT-GBD (OR: 0.43; CI 0.18 to 1.00; p=0.05; I2=66%), shorter lengths of hospital stay, with pooled standard mean difference of −2.53 (95% CI: −4.28 to −0.78; p=0.005, I2=98%); required significantly less re-interventions (OR: 0.16; CI: 0.04 to 0.042; p=0.0002; I2=32%) resulting in significantly less unplanned readmissions (OR: 0.16; 95% CI: 0.05 to 0,53; p=0.003, I2=79%).24 The findings of the current study are in-line with the all of the above studies. Furthermore, we were also able to show that the rates of recurrent acute cholecystitis were significantly less in the EUS-GBD group.
The performance of EUS-GBD also allows endoscopic removal of gallstones using standard endoscopic techniques and a stone clearance rate of 88% after a mean (SD) number of 1.25 (0.46) sessions of cholecystoscopy was reported.12 Since the largest diameter of stent is 15 mm, larger gallstones would not be spontaneously passed out to cause gallstone ileus. On the other hand, in patients that do not want a second endoscopic procedure, long-term placement of the metal stent is another option.9
Certainly, the most important concern to a surgeon when performing EUS-GBD is the risk of leakage from the draining organs. The integrity of a surgical anastomosis is dependent on several factors including surgical technique, blood supply, the type of organ and presence of tension. However, when performing EUS-guided anastomosis, the integrity of the anastomosis is mostly dependent on the stent that is placed. Hence, the use of LAMS is of paramount importance, as the lumen apposing force generated by LAMS is significantly higher than that of other types of stents.25 The safety of the EUS-guided anastomosis is confirmed by the results of the meta-analyses and a few large-scale studies.24 26 In fact, the occurrence of an anastomotic leak after successful placement of a LAMS for acute cholecystitis is exceedingly rare. Hence, we have seen a proliferation of reports on various EUS-guided anastomotic procedures including gastro-enterostomies, afferent limb drainage, bile ducts and pancreatic fluid collections in recent years.27–32
On the other hand, endoscopist performing EUS-GBD should be wary of the fact that after EUS-GBD, the difficulty of performing a cholecystectomy subsequently could be significantly increased. Thus, this procedure should only be offered to patients where cholecystectomy would not be considered in the future. This decision should best be made in the consensus of a multidisciplinary team involving surgeons, radiologist and endoscopist. The experience of cholecystectomy after EUS-GBD is currently very limited and only one study have reported the outcomes.33 Hence, it is too early to conclude whether prior EUS-GBD would or would not result in higher complications from cholecystectomy.
There are a number of limitations to the current study. First, EUS-GBD were all performed by high-volume institutions and the results may not be generalised to other centres. Second, the follow-up period of the current study is not adequate to address the long-term outcomes of EUS-GBD. However, given the demographics of study population, most will have a limited life expectancy to experience recurrent stones. Also, the current study did not compare the cost-effectiveness of the two procedures. EUS-GBD requires dedicated devices and expertise, and the cost of the initial cost of the procedure is higher than that of PT-GBD. However, given that PT-GBD is associated with significantly higher numbers of re-interventions and re-admissions, the cost of reinterventions and readmissions will be more. Hence, whether the two procedures are comparable in terms of cost-effectiveness is arguable. In addition, the assessment of quality of life measurements were not possible in this study, as many of the included patients were old and fragile and could not comprehend with the questionnaires. On the other hand, tube dislodgements accounted for the majority of the adverse events in the PT-GBD arm. This rate appeared to be higher than that reported in the literature that is between 16% to 25%.21 34 This may be due to the fact that in our group of patients, cholecystectomy could not be performed due to poor premorbid and PT-GBD was used as a definite treatment. In the CHOCOLATE trial, the rate of adverse events in the PT-GBD arm was also up to 65%.16 Among them, 45 out of 68 of the patients (66%) assigned to the PT-GBD group required additional percutaneous or surgical interventions related to cholecystitis. The findings again echoing the potential issues with PT-GBD as noted in this study. In addition, the estimated power of PT-GBD as a predictor to recurrent acute cholecystitis was 53.4%. This means that we only had that much chance to detect the true association if there was one. Low statistical power may also imply overestimation of OR and lower likelihood that the observed effect is a real one. Lastly, EUS-GBD was performed by a single type of lumen apposing stents in this study, and whether the results can be generalised to other types of stents is uncertain.
In conclusion, EUS-GBD improved outcomes compared to PT-GBD in patients suffering from acute cholecystitis who were not candidates for cholecystectomy. EUS-GBD should be the procedure of choice in these patients provided that the expertise is available after a multidisciplinary meeting. Further studies are required to determine the long-term efficacy.
Contributors AYBT was responsible for study design, data collection, analysis of data and writing of the manuscript.MK, MP-M, TO, SC, CS-H, SO, RT-Y and TT were responsible for data collection and analysis of data. TI was responsible for study design and data collection analysis of data. KTW was responsible for data collection, analysis of data and writing of the manuscript. C-HL was responsible for analysis of data and writing of the manuscript. PWYC, EKN and JL were responsible for study design, analysis of data and final approval of the manuscript.
Funding This study was supported by the American Society of Gastrointestinal Endoscopy research awards.
Competing interests AYBT is a consultant for Boston Scientific, Cook, Taewoong and Microtech Medical Corporations.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. Individual data that underlie the results reported in this article after deidentification.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.