Introduction

Acute cholecystitis (AC) accounts for 30% of emergency admissions to general surgery services and is the second most frequent cause of complicated intra-abdominal infection [1]. Currently, laparoscopic cholecystectomy (LC) is the gold standard in the treatment of AC. However, in patients with high surgical risk, comorbidities, or advanced age, LC is associated with high rates of morbidity (31%) and postoperative mortality (4%)—much higher than those obtained in patients with low surgical risk [2].

In patients with a level of surgical risk that outweighs the possible benefits, the use of non-surgical treatments has become widespread [3, 4]. The most commonly used alternative to surgery is percutaneous cholecystostomy (PC), which consists of the percutaneous puncture of the gallbladder and the placement of a drainage catheter [5].

The exponential increase in the use of PC over the last decade is due to two main factors [6]: population aging, which increases the numbers of patients with high surgical risk, and the publication of the Tokyo guidelines (TG) devised to standardize the diagnosis, management and treatment of acute lithiasic cholecystitis, and which recommend the use of PC in a large group of patients [7]. The TG have been rapidly accepted worldwide. Recently, however, the utility of PC versus LC in patients with high surgical risk has been questioned, and it has been suggested that PC may be overindicated [4, 8].

The theoretical advantages offered by PC are the rapid resolution of sepsis and the optimal preparation of the patient for elective LC [5]. Its main drawback is the possibility of recurrence of AC or other biliary events while awaiting LC.

Thus, many questions about PC remain unanswered: how should the catheter be handled and removed? When is the best time to perform LC? Should cholecystectomy be offered to all patients after PC? Does PC complicate subsequent cholecystectomy? How good is the adherence to the TG in real life with regard to the indication of PC? To answer these questions, we devised the present study involving patients undergoing PC at our center.

The main objective of the study is to assess the indications, patients’ characteristics and results with regard to morbidity and mortality in patients with acute lithiasic cholecystitis managed with PC and in patients who undergo cholecystectomy after PC. As secondary objectives, we evaluate adherence of the indication of PC to the Tokyo 18 guidelines, PC management, the adequate duration of drainage, and the rates of elective and emergency cholecystectomy after PC.

Material and methods

Study setting and patient identification

This retrospective observational study was conducted between January 2016 and February 2021. The study was authorized by the Ethics Committee (CEIM) of the public Hospital General Universitario de Alicante (HGUA: nº 2021-120, ISABIAL 0253). The requirement of informed consent was waived due to the retrospective nature of the study and the absence of any risk. Data on patients with PC in the study period were extracted from the HGUA database and analysed retrospectively.

Cohort selection and data collection

The inclusion criteria were: patients undergoing PC diagnosed with AC following the TG13 and TG18 diagnostic criteria [9, 10]. The exclusion criteria were: patients undergoing PC for causes other than AC, such as neoplasms, bile duct alterations or non-therapeutic diagnostic purposes; patients who had previously undergone endoscopic drainage.

The characteristics of the sample undergoing PC, main indications, evolution and clinical results were studied first. Subsequently, a retrospective analytical study was designed to compare various cohorts: elective or emergency surgery and management with PC alone; patients with/without a high surgical risk; and elective vs emergency surgery. Patients’ main characteristics, associated morbidity (complications according to Clavien-Dindo grade and 90-day mortality, need for new drain placement, and surgical approach (laparoscopic vs. open) were compared, following the STROCSS 2019 guidelines [11].

Variables were compiled from a review of the digitized medical histories which included one year of follow-up after use of PC or until death. The demographic variables studied were age and sex. Functional status was assessed according to the ASA scale and comorbidity using the Charlson Comorbidity Index (CCI) and chronic anticoagulant treatment with oral or injectable anticoagulants after diagnosis AC [12, 13]. The type of radiological test used in the diagnosis (ultrasound, computed tomography (CT), nuclear magnetic resonance, cholangioresonance or a combination of these) and laboratory tests (C-reactive protein and leukocyte count) were recorded. The initial imaging study was performed by ultrasound; if locally evolved acute cholecystitis was suspected, or if the ultrasound diagnosis was unclear, the study was completed with CT (available in 24 h) or MRI. Marked local inflammation was defined as gangrenous or emphysematous AC, biliary perforation/peritonitis, or perivesical abscess. Patients were classified according to the TG13/18 severity scale: Grade I (mild), Grade II (moderate) or Grade III (severe). The degree of adherence to the TG13/18 was taken into account in the indication of PC.

The main indications for PC, total length of hospitalization, time from admission to drainage placement, and drainage duration in days were recorded. The procedure was considered successful when the patient did not require a new drain or emergency surgery, did not die due to the infection, and could be discharged from hospital after drain removal. Even though PC is not a surgical intervention, the Clavien-Dindo complication scale was used.

The patients who underwent cholecystectomy and those placed on the surgical waiting list were recorded, as was the type of surgery (emergency vs elective). In the case of emergency surgery the reason for the intervention was also reported. Approach, conversion rate to open surgery, total length of hospitalization, time from PC to cholecystectomy (in days), and finally complications according to the Clavien-Dindo classification were assessed [14]. Ninety-day mortality rates of patients both after PC and after cholecystectomy were recorded, as well as the causes. As regards clinical evolution, readmission rates for biliary causes (AC, biliary colic, choledocholithiasis, cholangitis and/or pancreatitis) and other causes were reported, along with time until readmission and main reason.

Description of the procedure

For the diagnosis of AC, we apply the definition in the TG in 2007 with its subsequent revisions in 2013 and 2018 [8, 9]. The criteria for a definitive diagnosis of AC were the presence of one of the symptoms from each of the following groups: local signs of inflammation (Murphy’s sign, right upper abdominal quadrant, mass/pain/tenderness), systemic signs of inflammation (fever, elevated C-reactive protein, elevated white blood cell count) and imaging findings characteristic of acute cholecystitis.

After the general assessment (physical examination, complementary tests and clinical status) the surgical team decided whether to proceed with PC or perform emergency surgery. At our institution, AC is managed in accordance with the TG; however, some variability is inevitable, since the guidelines themselves offer various therapeutic options for the same patient depending on their characteristics, the severity of the condition and the criteria of the surgeon who makes the initial assessment of the patient.

The PC was put in place by interventional radiologists. The technique was performed under local anesthesia in aseptic conditions, guided by ultrasound or CT [5]. Ultrasound-guided transhepatic PC using the Seldinger technique was the usual procedure. In our protocol, the drain is withdrawn a week after its insertion. Initially it is closed for 48 h; if it is adequately tolerated, we perform a control ultrasound. If correct emptying of the gallbladder is observed, the drain can be removed; if not, if the patient does not adequately tolerate the closure of the drain, it is maintained for another week, a cholangiography is performed, and if it demonstrates the passage of contrast into the common bile duct the drain can now be removed, If the study does not confirm cystic-choledochal permeability or the presence of choledocholithiasis, endoscopic treatment using ERCP or transcholecystic impulsion is considered.

Emergency cholecystectomy was performed by the oncall surgical team. If a laparoscopic approach was chosen, it was carried out using the French technique, with dissection of Calot's triangle until the Strasberg critical view of safety was achieved. If an open approach was selected or if conversion from laparoscopic surgery proved necessary, it was carried out via right subcostal laparotomy.

Statistical study

Statistical analysis was performed with IBM-SPSS 25.0®. The quantitative variables were expressed as measures of central tendency with means and medians, together with dispersion measures such as standard deviation and statistical range. The qualitative variables were expressed in the form of proportions or percentages.

Qualitative variables were compared using contingency tables and the chi-squared test. In variables with small samples (i.e., fewer than five individuals in 25% of the cells) the Fisher exact test was performed. To compare quantitative variables, the Mann–Whitney U test was applied since none of the variables had a normal distribution and in some cases the sample size was smaller than 30. Results with levels of p < 0. 05 were considered statistically significant.

Results

During the study period, PC was performed in 195 patients who met the selection criteria. A mean of 37.4 procedures were performed annually (σ = 6.8); the number increased gradually over the period; from 26 procedures in 2016 to a maximum of 46 in 2020. The mean age was 74 years, and comorbidity rates were high: ASA III/IV (59.5%) and a mean CCI of 5.5. Table 1 shows the main demographic data, functional status and comorbidity indices.

Table 1 General characteristics of the sample
Full size table

Diagnostic tests

The clinical diagnosis of AC was confirmed radiologically by ultrasound alone in 109 patients (55.9%), by ultrasound and CT in 35 (18%), by CT alone in 43 (22.1%) and by cholangio-magnetic resonance imaging in 6 (3.1%).

The mean leukocyte count at admission was 15,484 leukocytes/mm3 (σ = 7666), with a range of 660–45,000 leukocytes/mm3. Upon admission, 50 patients (25.6%) presented normal values; 61 (31.3%) had more than 18,000 leukocytes/mm3 and 76 (39%) between 11,000 and 18,000 leukocytes/mm3. The mean C-reactive protein level at admission was 17.6 mg/dl (σ = 12.4).

Main features of AC

One hundred and eighty-six patients (95.4%) had acute lithiasic cholecystitis, and the remaining nine (4.6%) were alithiasic. Sixty-five patients (33.3%) presented marked local inflammation: perivesical/hepatic abscess was the most frequent, in 39 patients (20%), gangrenous AC in 16 (8.2%), perforation in eight (4.1%) and emphysematous AC in two (1%).

The severity of AC according to the TG [8, 9] was: Grade I, 18 patients (9.2%) with a CCI > 5 in 12 (66.7%); Grade II, 108 patients (55.4%) with a CCI > 5 in 56 (51.9%); and Grade III, 69 patients (35.4%) with a CCI > 5 in 34 (49.3%). Ninety-nine patients (50.8%) adhered to the TG. Stratifying the rate of adherence according to severity, it was 23.1% in grade II patients and 100% in grade III.

Percutaneous cholecystostomy: indications and clinical results

The most frequent reason for PC was high surgical risk (36.4%). Other indications were: symptoms > 72 h (16.4%), septic shock (13.9%), marked local inflammation with perivesical abscess (11.3%), presence/suspicion of bile duct alteration (9.2%), refusal to undergo surgery (5.6%), combination of symptoms > 72 h and suspected bile duct alteration (3.6%), other causes (2.1%) and non-response to medical treatment (1.5%). There was a significantly greater proportion of patients with ASA > 3 and a higher mean CCI among those with indications for PC due to high surgical risk and septic shock than among those with other indications (p < 0.001). The mean time from admission to drain placement was 3.91 days (σ = 11.4).

Table 2 compares the cohort of patients who underwent PC due to high surgical risk versus the rest of the cohort. High-risk surgical patients had significantly higher mean CCI (6.3 vs 5) and a higher proportion of ASA score ≥ 3 (82.3% vs 47.6%). The high-risk group presented a higher proportion of AC grade III. However, the PC-related complication rate and the emergency cholecystectomy rate were similar in both subgroups. High-risk patients also had a lower rate of elective cholecystectomy (31.5% vs 7%), but this was not associated with a higher rate of readmission for biliary causes.

Table 2 High risk patients and no high risk patients cohorts
Full size table

The reason for drain removal was improvement in clinical symptoms or laboratory tests or both in 82.6%, poor clinical evolution in 12.3%, and accidental drain withdrawal in 5.1%. The procedure was considered a success in 130 patients (66.7%). Twenty-four patients (12.3%) presented PC-related complications: nine Clavien-Dindo type I (4.6%), fourteen Clavien-Dindo type IIIa (7.2%) and one Clavien-Dindo type IIIb (0.5%). 15 patients (8.7%) presented major complications (Clavien-Dindo ≥ IIIa) (Table 3). Mean length of hospitalization was 17 days (σ = 17) with a median of 12 days.

Table 3 Complications related to PC
Full size table

A new drain was required in 43 patients (22%). Twenty-six new PC (six due to accidental removal and twenty due to poor evolution of the infection) were put in place using the usual procedure. In the other 17 patients, the replacement was performed by endoscopic ultrasound-guided transmural drainage with a Hot-Axios ™ prosthesis.

The mean duration of drainage was 10.7 days (σ = 9.7), median 8 days (range: 1–75). The readmission rate for biliopancreatic pathology was 28.2%; the most frequent reasons were new AC (8.7%) and uncomplicated biliary colic (7.7%). In all, 69% of readmissions for this cause occurred during the first three months. The general mortality rate after PC was 8.7% at 30 days, 14.4% at 90 days, and 21% at one year. The following causes of death in the first 90 days were recorded: Infectious disease associated with AC (six patients), acute pancreatitis (two), Clostridium difficile colitis, COVID-19 infection (one), cancer (prostate and hematological (four)), cerebrovascular accident (two), cardiac surgery or progressive deterioration.

Cholecystectomy

Fifty-three patients (27.2%) underwent cholecystectomy: scheduled cholecystectomy was performed after PC in forty-four patients (22.6%), and emergency surgery was performed in nine cases (4.6%), in three due to poor evolution of the infection, in two due to the impossibility of removing the drain, in two due to a new AC after removal of the drain, and in the last two due to persistent biliary colic. Eight more patients are currently on the waiting list for the intervention. The mean time from drain placement to surgery was 189.6 days (σ = 158.2), median 149 days (range: 8–619 days). Elective surgery was always carried out more than eight weeks after the PC.

The surgical approach was laparoscopic in 35 patients (66%) and open in 18 34%): seven (13.2%) from the outset, and 11 (20.8%) were converted to open surgery after starting via laparoscopy. Eleven patients (20.8%) presented postoperative complications, five of them (9.4%) major (i.e., Clavien-Dindo ≥ IIIa). The mean hospital stay in the cholecystectomized patients was 4.2 days (σ = 5.3).

Table 4 compares the cohort of patients who underwent emergency/elective cholecystectomy versus those who did not undergo surgery. The no surgery group were older (77.91 years vs 66.87 years), had a higher mean CCI (6.25 vs 3.66), higher proportions of ASA scores ≥ 3 (67.6%% vs 39.6%) and of AC grade III (44.8% vs 13.2%). The differences were statistically significant in all cases.

Table 4 Surgical and non-surgical patient cohorts
Full size table

The indication for PC in the no surgery group was high surgical risk, which was significantly higher than in the surgery group. In addition, this group had a higher rate of readmission due to biliary events (21.6% vs 1.9%). No significant differences were found in terms of the 90-day mortality rate.

Table 5 compares patients undergoing emergency and elective cholecystectomy. No significant differences were found with regard to sex, mean age, CCI and proportion of ASA ≥ 3 between the two subgroups. Among our main findings regarding cholecystectomy after PC, the emergency surgery cohort presented significantly higher rates of initial open surgery (33.3% vs 9.1%) and conversion to open surgery (44.4% vs 15.9%) than the elective surgery cohort. Emergency surgery patients also presented longer hospital stay (7.78 days vs 3.43 days) although the differences were not significant. No differences were found in 90-day mortality or in the complication rate.

Table 5 Elective and emergency surgery cohorts
Full size table

Discussion

The treatment of AC has evolved over recent years, both due to the publication of guidelines that have standardized its management and due to technological advances in surgery and the establishment of the laparoscopic approach as the gold standard. Among other changes, the exponential increase in the use of PC is a result of the progressive aging and fragility of the population in whom emergency surgery represents too great risk. Against this background, this retrospective study was designed to compare our clinical results with those described in the literature.

PC is usually applied in patients with AC who are unable to undergo surgery due to comorbidities and/or shock or severe sepsis [15]. The guidelines also admit other indications for PC such as AC > 72 h, marked local inflammation or leukocyte count > 18,000 L/mm3 [3, 4]. The most frequent indications of PC in our series were high surgical risk and comorbidities (ASA ≥ 3, and CCI), evolution > 72 h, septic shock, and the presence/suspicion of bile duct alteration. This last indication is not usually cited in the guidelines, but the coexistence of biliary/pancreatic pathology suggests the presence of a serious inflammatory/infectious process for which PC may be a valid therapeutic option [16]..

The publication of the TG has had a great impact on the treatment of AC. This is especially true with regard to the classification of the condition into three grades, since this subdivision can act as a predictor of mortality, length of hospital stay, and conversion rate from laparoscopic to open surgery [9, 16, 17] Lin et al. concluded that the TG provide satisfactory and cost-effective clinical results [18]. In our series, we initially used the TG13 classification and later the TG18. We found that our severity rates are similar to those reported in other retrospective studies of PC [19].

PC is a procedure with a high technical success rate. It is safe, associated with low morbidity, effective, and allows rapid control of the focus of infection [16, 20]. Previous studies report clinical and technical success rates close to 90% [21]; applying strict criteria, we achieved a high success rate. Emergency cholecystectomy was only necessary in a low number of the patients who received PC (due to poor evolution, inability to remove PC, or recurrent AC after catheter removal), a rate lower than reported by Szabolcs et al. [22, 23]. Procedure-related mortality associated with PC ranges between 0 and 4% [24], in our series no deaths were reported. The systematic review by Winbladh et al. reports a complication rate of 14.8% [21]; our complication rate and the presence of Clavien-Dindo grade ≥ IIIA were lower. The mortality rate reported in patients with AC treated with PC ranges between 4.3% and 22.6%; however, it is attributable not to PC, but to the underlying severity of the condition and/or the presence of multiple pathologies [25,26,27]. Our 90-day mortality rate in patients treated with PC was similar to other series.

Around a fifth of our patients required the placement of a new PC, a figure similar to the 23% reported by Horn et al. [28]. In some cases, the new tube was inserted using endoscopic ultrasound-guided transmural drainage. This technique presents success rates approaching 80% [29], and it is especially useful in high surgical risk patients who are not candidates for elective cholecystectomy [30]. After withdrawal of the PC, the readmission rate for biliopancreatic disease was high. Similar rates of readmission have been described in the literature (20–40%).

Another ongoing debate is whether initial cholecystectomy is superior to PC in severe patients with AC. The systematic review by Ambe et al. [31] compared patients with PC vs early cholecystectomy, finding higher mortality, longer hospital stay and more readmissions for biliary pathology in patients with PC. However, La Greca et al. did not observe these differences in ASA 3 patients [32].

The CHOCOLATE randomized study comparing high surgical risk patients treated with PC or LC did not find differences in mortality rates, but observed higher rates of complications, reoperations, and recurrence of biliary pathology in PC patients [33]. However, we noted several relevant differences. In the CHOCOLATE study, the readmission rate for AC and biliary disease was 53% in the 68 patients included in the PC arm. In our study, with a larger sample, this rate was 28.2%. Furthermore, in the rate of complications associated with the use of PC, the authors of the CHOCOLATE study add the recurrence of biliary pathology, and so the rate of complications is very high (65%). The rate of emergency cholecystectomy after PC in our case was 4.6%, while in the CHOCOLATE study it was 16%. In addition, the rate of bile duct injury during cholecystectomy in the surgery group was 6% versus 3% in the PC group, with no statistically significant differences. Although the results were not comparable due to the study design, in our view the usefulness of PC lies in its rapid control of septic symptoms and the possibility of preparing some patients for a subsequent elective cholecystectomy.

One of the most controversial issues regarding PC is the duration of the placement of the drain. Some authors recommend keeping it in place until surgery or for a minimum of six weeks, since early removal is associated with complications. Others recommend its withdrawal when the AC has resolved, thus avoiding the need for new procedures due to malfunction or associated complications, since not all patients are candidates for surgery [34].

There is no consensus in the literature when the best time to perform LC is [35]. A retrospective study comparing 1211 patients with early cholecystectomy (0–8 weeks) versus late cholecystectomy (> 8 weeks) found that the early cholecystectomy group had a higher risk of complications and longer hospital stay [36]. In a meta-analysis published in 2022 Kourunis et al. [37] compared the performance of cholecystectomy during the first 30 days and at a later date in 41 prospective and retrospective studies, without finding statistically significant differences in the clinical results. However, there was a slightly higher conversion rate in the cholecystectomy group after 30 days: 7.2% vs. 8.3%.

The main guidelines recommend interval LC after PC withdrawal [4, 8]. There are many reasons why not all patients undergo surgery after PC: among them, the most important is their frailty and the high mortality due to causes other than biliary events [38], [39]. In our sample, the cohort of patients who underwent surgery had significantly lower mean ages and significantly better baseline situations. For this reason, cholecystectomy should be offered to all patients fit for surgery, but it is necessary to prioritize its indication and follow-up in patients with a higher risk of recurrence of biliary events and with an acceptable surgical risk.

It has been postulated that the inflammation caused by PC may increase the technical difficulty during cholecystectomy. In a retrospective analysis of elective cholecystectomy in patients treated with or without PC, patients treated with PC were assigned to open surgery more frequently [40]. In our series, emergency cholecystectomy presented a significantly higher conversion rate than elective surgery.

The strengths of the study are the prompt implementation of the TG which helped to standardize the indications between surgeons, the high sample size, and the long inclusion period. The main limitation of the study is its retrospective nature. This means that there is a risk of information bias, due to differences in the criteria for indicating PC or surgery, technical aspects during surgery, or the length of follow-up. The information available may also contain errors.

The Grading of Recommendation Assessment, Development and Evaluation (GRADE) [41] consider observational studies as low quality evidence. Accordingly, we cannot offer recommendations regarding of the major controversies. However, from this point of view, we can propose future research suggestions, a consensus conference between experts or an experts´ consensus using the policy of Delphi method to offer recommendations of controversies such as how the catheter should be handled or when cholecystectomy should be performed after PC.

Conclusion

PC achieves improvements in the inflammation and infection associated with AC. Its overall success rate in these habitually frail patients was 66.66%, and only 4.6% required emergency cholecystectomy. The rate of recurrence of AC and other biliary events is high after PC, and so it cannot be considered a definitive treatment. In our series, prior PC did not necessarily imply an increase in technical difficulty during laparoscopic cholecystectomy.

Adherence to the TG in our series was no higher than 50%, due probably to the early indication of PC in grade II ACs. The TG do not give specific recommendations regarding the management of PC or the indication of cholecystectomy after PC. Randomized studies should now be carried out to shed more light on these two issues.