Early laparoscopic cholecystectomy after percutaneous transhepatic gallbladder drainage in acute cholecystitis is appropriate and safe: an inverse probability of treatment weighting analysis

Abstract

Background

Acute cholecystitis (AC) is a common emergency requiring timely surgical intervention. While laparoscopic cholecystectomy (LC) is the standard treatment, the optimal timing for LC following percutaneous transhepatic gallbladder drainage (PTGBD) remains debated. This study evaluates and compares the safety and efficacy of early LC after PTGBD, immediate LC without PTGBD, and delayed LC following PTGBD.

Methods

A retrospective cohort study was conducted at a Level I referral center, analyzing 1436 patients diagnosed with AC and managed surgically between 2010 and 2018. Patients were categorized into three groups: early LC after PTGBD (Early group, n = 18), immediate LC without PTGBD (Immediate group, n = 1243), and delayed LC following PTGBD (Delayed group, n = 175). Patient demographics, clinical characteristics, and surgical outcomes were analyzed using inverse probability of treatment weighting (IPTW) to adjust for baseline differences.

Results

Compared to the Immediate group, the Early group had a higher proportion of older patients and multiple comorbidities. After adjustment, adverse event rates were similar between both groups, but major complications were lower in the Early group. Compared to the Delayed group, the Early group demonstrated significantly shorter hospital stays and lower major complication rates.

Conclusions

Early LC after PTGBD during the same admission is a viable alternative to delayed LC, reducing hospital stay and complications. Immediate LC remains the preferred approach for eligible patients without prior PTGBD. Further prospective studies are needed to refine the optimal timing of LC following PTGBD.

Introduction

Acute cholecystitis (AC) is a frequently seen condition in routine clinical practice and is a common cause of acute abdomen. Urgent treatment may be necessary when complications like gangrenous cholecystitis, emphysematous cholecystitis, or gallbladder torsion arise [1–3]. Surgery is recognized as the definitive treatment option. In the past, open cholecystectomy (OC) was the conventional standard for treating AC. With advancements in laparoscopic equipment and growing experience in minimally invasive surgery (MIS), which have enhanced safety and demonstrated benefits in reducing complications, laparoscopic cholecystectomy (LC) has become the preferred surgical approach and is now the first-line treatment for AC [3–5].

The Tokyo Guidelines, first published in 2007, served as the first globally recognized guidelines for diagnosing and managing AC [6]. They were subsequently revised, resulting in the updated versions known as the Tokyo Guidelines 2013 (TG13) and Tokyo Guidelines 2018 (TG18) [2, 7]. The TG13 recommended that the diagnostic criteria for AC should be based on a combination of local signs, systemic signs of inflammation, and imaging findings. Due to its high sensitivity and specificity, TG18 continued to adopt the diagnostic framework established by TG13 (Table 1). In addition, TG13 classified acute cholecystitis into three grades, a classification system that has been shown to be a valuable tool for predicting patient prognosis [8]. Therefore, TG18 retained the same grading system without modification (Table 2). In the TG18, an updated treatment flowchart for AC was presented. For patients with Grade I or Grade II AC, early LC is recommended if they fulfill the criteria of Charlson Comorbidity Index (CCI) ≤ 5 and American Society of Anesthesiologists physical status classification (ASA-PS) ≤ 2. However, for Grade II patients, the surgery should be carried out by experienced surgeons. In certain advanced centers with specialized surgical teams, selected Grade III AC patients who meet specific strict criteria may also undergo immediate LC. For patients deemed unsuitable for early surgery, the TG18 advises performing early or urgent percutaneous transhepatic gallbladder drainage (PTGBD), followed by delayed LC once the patient’s condition stabilizes [3].

Table 1.

Diagnostic criteria of acute cholecystitis by the Tokyo guidelines 2013/2018

A. Local signs of inflammation etc.:

1. Murphy’s sign 2. Mass/pain/tenderness over the right upper quadrant (RUQ) of the abdomen

B. Systemic signs of inflammation etc.:

1. Elevated white blood cell (WBC) count 2. Elevated C-reactive protein (CRP) 3. Fever

C. Imaging findings

Suspected diagnosis: one item in A + one item in B

Definite diagnosis: one item in A + one item in B + C

Table 2.

Severity and classification of acute cholecystitis by the Tokyo guidelines 2018

Grade I (mild) acute cholecystitis:

1. Acute cholecystitis in a healthy patient without organ dysfunction 2. Inflammatory changes in the gallbladder are mild, making cholecystectomy a safe and low-risk procedure 3. Acute cholecystitis that does not meet the criteria for grade III or grade II cholecystitis

Grade II (moderate) acute cholecystitis, associated with any one of the following conditions:

1. Elevated WBC count (> 18,000/mm3) 2. Palpable tender mass in the right upper abdominal quadrant 3. Duration of complaints > 72 h 4. Marked local inflammation (gangrenous cholecystitis, pericholecystic abscess, hepatic abscess, biliary peritonitis, emphysematous cholecystitis)

Grade III (severe) acute cholecystitis, associated with dysfunction of any one of the following organs/systems:

1. Cardiovascular dysfunction: hypotension requiring treatment with dopamine ≥ 5 µg/kg per min, or any dose of norepinephrine 2. Neurological dysfunction: decreased level of consciousness 3. Respiratory dysfunction: PaO2/FiO2 ratio < 300 4. Renal dysfunction: oliguria, creatinine > 2.0 mg/dl 5. Hepatic dysfunction: PT-INR > 1.5 6. Hematological dysfunction: platelet count < 100,000/mm3

In critically ill patients with AC, the risk associated with anesthesia is expected to rise if immediate surgical intervention is performed. In contrast, PTGBD is considered an effective alternative for rapidly alleviating symptoms of acute cholecystitis [9]. However, despite numerous published studies with varying conclusions, the ideal timing for cholecystectomy after PTGBD remains a topic of debate [3]. In clinical practice, studies have reported mixed outcomes in high-risk patients undergoing PTGBD followed by early or urgent surgery, including prolonged operative times and increased intraoperative bleeding [10, 11]. However, one study found that there was no considerable difference between the two strategies [12]. Additionally, two studies reported favorable outcomes when LC was performed 4 to 6 weeks after PTGBD, in terms of reduced bleeding volume, shorter operating times, lower conversion rates to OC, and fewer complications [13, 14]. However, prolonged indwelling of an external drain can have a significant negative impact on patients’ quality of life (QoL) and contribute to increased medical costs [3, 15, 16]. Therefore, the present study aimed to evaluate and compare the safety and efficacy of early LC after PTGBD, immediate LC without PTGBD, and delayed LC following PTGBD.

Materials and methods

Study patients

The retrospective cohort study included patients from a single level I referral center, which Handles approximately 300 medical patient visits to the emergency department (ED) daily and performs surgery on at least 1500 patients with acute abdominal conditions annually. We retrospectively reviewed our acute abdomen database and the medical records of patients who were admitted to the ED of our hospital due to AC and underwent either PTGBD or surgical management between September 2010 and July 2018. Using the TG13/TG18 diagnostic criteria, AC was diagnosed based on clinical symptoms, blood test results, and imaging findings. Computed tomography (CT) was performed for all patients to confirm the diagnosis. Patients younger than 18 years old, those who received conservative treatment, or those who underwent laparotomy were excluded from the study. After follow-up, patients who underwent PTGBD followed by delayed LC were also included, while those received PTGBD without subsequent surgical management were ultimately excluded. A total of 1436 patients were included for analysis. Among these patients, 18 (1.25%) underwent PTGBD followed by LC during the same hospitalization (Early group), while 1243 patients (86.56%) underwent emergency LC without prior PTGBD (Immediate group). The remaining 175 patients (12.19%) underwent scheduled LC following PTGBD and were categorized as the Delayed group.

Methods

At our institute, LC is routinely performed by experienced, certified hepatobiliary surgeons or acute care surgeons, each with a cumulative surgical experience of more than 300 cases, using either the three-port or four-port technique. In all cases, the decision to convert from LC to open surgery was made at the discretion of the operating surgeons. Major surgical complications, whether identified intraoperatively or diagnosed postoperatively, included bile duct injuries, intraoperative or postoperative massive bleeding greater than 500 mL, hollow organ perforations, and other serious conditions. With improvements in surgical instruments, techniques, and increasing experience in MIS, some of these complications could potentially be managed laparoscopically, thereby avoiding the need for conversion to open surgery. In this study, adverse events were defined as the occurrence of major complications or conversion to open surgery.

Data collection and statistical analysis

Information on demographics, characteristics, and surgical outcomes were collected from medical records. Elevated liver function tests (LFT) were defined as AST (Aspartate aminotransferase) or ALT (Alanine aminotransferase) levels greater than twice the normal upper limit, while hyperbilirubinemia was defined as serum direct bilirubin > 1 mg/dL or total bilirubin > 3 mg/dL. Total hospital stay and ICU stay were defined as the overall duration of hospitalization for the primary management of acute cholecystitis (AC) and any subsequent complications, including readmission due to PTGBD dislodgement or recurrent infections. Continuous data were presented as median with interquartile range (IQR) and mean ± standard error of the mean (SEM), and were analyzed using the Mann–Whitney U test and t-test, respectively. Categorical variables were presented as percentages, and Pearson’s χ² test or Fisher’s exact test was used to determine statistically significant differences between groups. To adjust for differences in patient characteristics between groups, inverse probability of treatment weighting (IPTW) analysis using propensity scores was performed to compare surgical outcomes [17]. IPTW is a valuable statistical tool in medical research, particularly in observational studies where randomization is not possible. By weighting individuals based on their probability of receiving a treatment, IPTW helps balance covariates and improve the validity of causal inferences, making observational studies more comparable to randomized controlled trials. A key advantage of IPTW is its capacity to reduce confounding bias while retaining the entire study population, in contrast to propensity score matching, which may exclude unmatched cases [18]. This characteristic is particularly important when the available sample size is limited [19]. Even in smaller cohorts, IPTW can effectively adjust for measured confounders and achieve balance in baseline covariates between treatment groups, thereby mitigating the selection bias inherent in retrospective studies [20]. The propensity score was developed using a logistic regression model to estimate the probability of being assigned to the Early group compared with the Immediate group, as well as the probability of being assigned to the Early group compared with the Delayed group [3, 7]. Relevant covariates were carefully selected based on known or potential predictors of unfavorable clinical outcomes in patients with acute abdomen, as identified in previous studies. These covariates included factors such as age, comorbidities, TG13/18 severity grading, duration of symptom onset, and vital signs upon hospital arrival. All selected covariates were incorporated into the propensity model [3, 9, 20, 21], and patients with missing data for any of these covariates were excluded from the propensity score calculation. A p value < 0.05 was considered statistically significant. All statistical analyses were performed using the statistical software SPSS version 27.0 (IBM Corp., Armonk, NY, USA).

Results

Study characteristics: early group vs. immediate group

A total of 18 patients were included in the Early group, and 1243 patients were included in the Immediate group. Patients in the Early group were significantly older than those in the Immediate group, with a median age of 65 years compared to 52 years (p = 0.003) and the mean age of 63.7 ± 3.49 years vs. 52.6 ± 0.43 years, p = 0.002). Additionally, a higher proportion of patients in the Early group were over 70 years old (38.9% vs. 14.2%, p = 0.010). The Early group had a significantly higher proportion of patients with multiple comorbidities (16.7% vs. 2.0%, p = 0.006). There were no significant differences between the groups in terms of liver cirrhosis (5.6% vs. 1.0%, p = 0.183) or history of prior abdominal surgery (22.2% vs. 17.7%, p = 0.544).

Regarding clinical presentation, a significantly higher proportion of patients in the Early group presented with fever compared to the Immediate group (55.6% vs. 16.0%, p < 0.001). The two groups showed no significant differences in the proportion of patients with symptom onset exceeding 72 h (p = 1.000) or shock on admission (SBP < 90 mmHg) (p = 1.000). The leukocyte count tended to be higher in the Early group compared to the Immediate group, though the difference did not reach statistical significance (median: p = 0.050; mean: p = 0.053). Platelet counts and PT-INR (international normalized ratio of prothrombin time) values also showed no significant differences between the two groups. Additionally, there were no significant differences between groups in the proportion of patients with elevated liver function tests (LFT) (p = 0.383) or hyperbilirubinemia (p = 1.000).

In terms of disease severity, the Early group had a higher proportion of Grade II patients compared to the Immediate group (55.5% vs. 31.3%), while the Immediate group had a significantly higher proportion of Grade I cases (66.2% vs. 38.9%). The proportion of Grade III cases was relatively low in both groups (5.6% in the Early group vs. 2.5% in the Immediate group). The time from admission to surgery was significantly longer in the Early group, with a median of 6.3 days compared to 0.9 days in the Immediate group (p < 0.001), and a mean of 7.0 ± 0.72 days compared to 1.9 ± 0.06 days, respectively (p < 0.001). In the Early group, the median interval from PTGBD to surgery was 5.8 days, with a mean interval of 6.5 ± 0.66 days. Besides, the proportion of patients undergoing ERCP (endoscopic retrograde cholangiopancreatography) did not differ significantly between the Early and Immediate groups (5.6% vs. 14.6%, p = 0.497) (Table 3).

Table 3.

Clinical characteristics and demographics of patients underwent post-PTGBD early laparoscopic cholecystectomy (Early group) and immediate laparoscopic cholecystectomy (Immediate group)

	Early	Immediate	p value
	N = 18	N = 1243
Age (years)
Median (IQR)	65 (53–78)	52 (41–64)	0.003
Mean ± SEM	63.7 ± 3.49	52.6 ± 0.43	0.002
> 70	7 (38.9%)	177 (14.2%)	0.010
Gender			0.812
Male	9 (50.0%)	685 (55.1%)
Female	9 (50.0%)	558 (44.9%)
Liver cirrhosis	1 (5.6%)	13 (1.0%)	0.183
≥ 3 Comorbidities	3 (16.7%)	25 (2.0%)	0.006
Abdominal surgical history	4 (22.2%)	220 (17.7%)	0.544
Onset > 72 h	4 (22.2%)	282 (22.7%)	1.000
Fever (BT > 38°)	10 (55.6%)	199 (16.0%)	< 0.001
SBP on admission
Median (IQR)	154 (131–181)	143 (125–161)	0.167
Mean ± SEM	155 ± 7.1	145 ± 0.8	0.133
Shock (< 90 mmHg)	0 (0.0%)	11 (0.9%)	1.000
Leukocyte (103/µL)
Median (IQR)	13.2 (10.5–17.5)	11.1 (8.2–14.6)	0.050
Mean ± SEM	14.1 ± 1.24	11.8 ± 1.41	0.053
Platelet count (103/µL)
Median (IQR)	226.5 (187.0–258.5)	233.0 (193.0–280.0)	0.568
Mean ± SEM	230.5 ± 14.57	243.8 ± 2.92	0.583
PT-INR
Median (IQR)	1.10 (1.10–1.20)	1.10 (1.00–1.20)	0.134
Mean ± SEM	1.14 ± 0.021	1.12 ± 0.005	0.481
Elevated LFT > 2 times	2 (12.5%, n = 16)	287 (24.4%, n = 1174)	0.383
Hyperbilirubinemia	2 (11.8%, n = 17)	156 (13.3%, n = 1175)	1.000
Tokyo Guidelines 2018			0.037
Grade 1	7 (38.9%)	823 (66.2%)
Grade 2	10 (55.5%)	389 (31.3%)
Grade 3	1 (5.6%)	31 (2.5%)
Application of ERCP	1 (5.6%)	181 (14.6%)	0.497
Interval before operation
Post-PTGBD (day)
Median (IQR)	5.8 (4.2–8.1)
Mean ± SEM	6.5 ± 0.66
Post-admission (day)
Median (IQR)	6.3 (4.8–9.1)	0.9 (0.6–2.5)	< 0.001
Mean ± SEM	7.0 ± 0.72	1.9 ± 0.06	< 0.001

*PTGBD, percutaneous transhepatic gallbladder drainage; IQR, interquartile range; SEM, standard error of the mean; SBP, systolic blood pressure; PT-INR, international normalized ratio of prothrombin time; LFT, liver function tests; ERCP, endoscopic retrograde cholangiopancreatography

The results or descriptions highlighted in bold are solely intended to facilitate readers in readily identifying the key points in the tables

Study characteristics: early group vs. delayed group

A total of 175 patients were included in the Delayed group. There was no significant difference in age between the two groups, with a median age of 65 years in the Early group compared to 69 years in the Delayed group (p = 0.493) and a mean age of 63.7 ± 3.49 years vs. 66.3 ± 1.10 years (p = 0.461). The proportion of patients aged over 70 years was also comparable between the Early and Delayed groups (38.9% vs. 46.9%, p = 0.623). There were no significant differences between the two groups in terms of liver cirrhosis (5.6% vs. 1.7%, p = 0.326), history of prior abdominal surgery (22.2% vs. 17.1%, p = 0.529), or the proportion of patients with three or more comorbidities (16.7% vs. 13.7%, p = 0.722).

Regarding clinical presentation, a significantly higher proportion of patients in the Early group presented with fever compared to the Delayed group (55.6% vs. 29.1%, p = 0.032). The two groups showed no significant differences in the proportion of patients with symptom onset exceeding 72 h (p = 1.000) or shock on admission (p = 1.000). Leukocyte counts were similar between the two groups (p = 0.734 for median; p = 0.729 for mean). However, platelet counts were significantly lower in the Delayed group (median: 192.0 × 10³/µL vs. 226.5 × 10³/µL, p = 0.028; mean: 194.1 ± 4.88 × 10³/µL vs. 230.5 ± 14.57 × 10³/µL, p = 0.023). PT-INR values showed no significant differences between the two groups. Additionally, there were no significant differences between groups in the proportion of patients with elevated LFT > 2 times (p = 1.000) or hyperbilirubinemia (p = 1.000).

With respect to disease severity, the distribution of TG13/TG18 severity grades was similar between the two groups (p = 0.464), except for a slightly lower proportion of Grade III cases in the Early group compared to the Delayed group (5.6% vs. 16.6%). The time from PTGBD to surgery was, unsurprisingly, significantly shorter in the Early group, with a median of 5.8 days compared to 73.0 days in the Delayed group (p < 0.001), and a mean of 6.5 ± 0.66 days versus 129.3 ± 18.49 days (p = 0.035). Concerning the application of ERCP, there was no significant difference between the Early and Delayed groups, with rates of 5.6% and 6.9%, respectively (p = 1.000) (Table 4).

Table 4.

Clinical characteristics and demographics of patients underwent post-PTGBD early laparoscopic cholecystectomy (Early group) and post-PTGBD delayed laparoscopic cholecystectomy (Delayed group)

	Early	Delayed	p value
	N = 18	N = 175
Age (years)
Median (IQR)	65 (53–78)	69 (57–77)	0.493
Mean ± SEM	63.7 ± 3.49	66.3 ± 1.10	0.461
> 70	7 (38.9%)	82 (46.9%)	0.623
Gender			0.805
Male	9 (50.0%)	96 (54.9%)
Female	9 (50.0%)	79 (45.1%)
Liver cirrhosis	1 (5.6%)	3 (1.7%)	0.326
≥ 3 Comorbidities	3 (16.7%)	24 (13.7%)	0.722
Abdominal surgical history	4 (22.2%)	30 (17.1%)	0.529
Onset > 72 h	4 (22.2%)	41 (23.4%)	1.000
Fever (BT > 38°)	10 (55.6%)	51 (29.1%)	0.032
SBP on admission
Median (IQR)	154 (131–181)	144 (123–164)	0.207
Mean ± SEM	155 ± 7.1	145 ± 2.2	0.186
Shock (< 90 mmHg)	0 (0.0%)	2 (1.1%)	1.000
Leukocyte (103/µL)
Median (IQR)	13.2 (10.5–17.5)	14.4 (10.4–17.7)	0.734
Mean ± SEM	14.1 ± 1.24	14.5 ± 0.42	0.729
Platelet count (103/µL)
Median (IQR)	226.5 (187.0–258.5)	192.0 (142.0–239.0)	0.028
Mean ± SEM	230.5 ± 14.57	194.1 ± 4.88	0.023
PT-INR
Median (IQR)	1.10 (1.10–1.20)	1.20 (1.10–1.30)	0.143
Mean ± SEM	1.14 ± 0.021	1.20 ± 0.013	0.153
Elevated LFT > 2 times	2 (12.5%, n = 16)	26 (15.9%, n = 164)	1.000
Hyperbilirubinemia	2 (11.8%, n = 17)	23 (14.4%, n = 160)	1.000
Tokyo Guidelines 2018			0.464
Grade 1	7 (38.9%)	54 (30.8%)
Grade 2	10 (55.5%)	92 (52.6%)
Grade 3	1 (5.6%)	29 (16.6%)
Application of ERCP	1 (5.6%)	12 (6.9%)	1.000
Interval before operation
Post-PTGBD (day)
Median (IQR)	5.8 (4.2–8.1)	73.0 (56.0–105.0)	< 0.001
Mean ± SEM	6.5 ± 0.66	129.3 ± 18.49	0.035

*PTGBD, percutaneous transhepatic gallbladder drainage; IQR, interquartile range; SEM, standard error of the mean; SBP, systolic blood pressure; PT-INR, international normalized ratio of prothrombin time; LFT, liver function tests; ERCP, endoscopic retrograde cholangiopancreatography

The results or descriptions highlighted in bold are solely intended to facilitate readers in readily identifying the key points in the tables

Outcome evaluation

Tables 5 and 6 summarize the outcomes comparing the Early group with the Immediate and Delayed groups. In the Early vs. Immediate comparison, unadjusted analysis showed a higher overall adverse event rate in the Early group (11.1% vs. 5.7%), though not statistically significant (p = 0.280). After IPTW adjustment, rates were similar (4.9% vs. 5.6%, p = 0.506). Conversion to laparotomy did not differ significantly in either unadjusted (11.1% vs. 4.7%, p = 0.210) or adjusted analyses (4.9% vs. 4.8%, p = 0.922). Major complications were significantly higher in the Immediate group after adjustment (1.5% vs. 0.0%, p < 0.001).

Table 5.

Outcomes of patients underwent post-PTGBD early laparoscopic cholecystectomy (Early group) and immediate laparoscopic cholecystectomy (Immediate group)

	Unadjusted			Adjusted by IPTW (frequency)
	Early (N = 18)	Immediate (N = 1243)	p value	Early	Immediate	p value
Adverse events	2 (11.1%)	71 (5.7%)	0.280	54 (4.9%)	63 (5.6%)	0.506
Conversion to laparotomy	2 (11.1%)	58 (4.7%)	0.210	54 (4.9%)	54 (4.8%)	0.922
Major complication	0 (0.0%)	23 (1.0%)#	1.000	0 (0.0%)	17 (1.5%)	< 0.001
Total hospital stay (day)
Median (IQR)	12.0 (8.9–14.8)	4.0 (3.0–6.0)	< 0.001	11.0 (5.0–17.0)	4.0 (3.0–6.0)	< 0.001
Mean ± SEM	13.2 ± 1.66	4.9 ± 0.10	< 0.001	11.6 ± 0.18	4.9 ± 0.11	< 0.001
Total ICU stay (day)
Median (IQR)	0.0 (0.0–0.0)	0.0 (0.0–0.0)	0.005	0.0 (0.0–0.0)	0.0 (0.0–0.0)	0.056
Mean ± SEM	0.6 ± 0.35	0.1 ± 0.05	0.328	0.1 ± 0.02	0.1 ± 0.06	0.086

^*PTGBD, percutaneous transhepatic gallbladder drainage; IPTW, inverse probability of treatment weighting; IQR, interquartile range; SEM, standard error of the mean

^#Ten patients with bile duct injury, 4 with bleeding, and 9 with others

The results or descriptions highlighted in bold are solely intended to facilitate readers in readily identifying the key points in the tables

Table 6.

Outcomes of patients underwent post-PTGBD early laparoscopic cholecystectomy (Early group) and post-PTGBD delayed laparoscopic cholecystectomy (Delayed group)

	Unadjusted			Adjusted by IPTW (frequency)
	Early (N = 18)	Delayed (N = 175)	p value	Early	Delayed	p value
Adverse events	2 (11.1%)	28 (16.0%)	0.744	13 (7.3%)	22 (13.7%)	0.073
Conversion to laparotomy	2 (11.1%)	20 (11.4%)	1.000	13 (7.3%)	17 (10.6%)	0.340
Major complication	0 (0.0%)	19 (10.9%)#	0.225	0 (0.0%)	15 (9.3%)	< 0.001
Total hospital stay (day)
Median (IQR)	12.0 (8.9–14.8)	13.0 (11.0–17.0)	0.238	10.0 (5.0–12.0)	12.0 (10.0–17.0)	< 0.001
Mean ± SEM	13.2 ± 1.66	15.4 ± 0.74	0.364	11.0 ± 0.50	15.0 ± 0.76	< 0.001
Total ICU stay (day)
Median (IQR)	0.0 (0.0–0.0)	0.0 (0.0–0.0)	0.680	0.0 (0.0–0.0)	0.0 (0.0–0.0)	0.180
Mean ± SEM	0.6 ± 0.35	0.6 ± 0.14	0.932	0.3 ± 0.09	0.5 ± 0.13	0.144

^*PTGBD, percutaneous transhepatic gallbladder drainage; IPTW, inverse probability of treatment weighting; IQR, interquartile range; SEM, standard error of the mean

^#Six patients with bile duct injury, one with duodenal perforation, 3 with bleeding, and 9 with others

When comparing the Early and Delayed groups, overall adverse events were slightly lower in the Early group, but differences were not significant (11.1% vs. 16.0%, p = 0.744; adjusted: 7.3% vs. 13.7%, p = 0.073). Conversion rates were similar in both unadjusted (11.1% vs. 11.4%, p = 1.000) and adjusted analyses (7.3% vs. 10.6%, p = 0.340). As with the Immediate group comparison, major complications were significantly higher in the Delayed group after adjustment (9.3% vs. 0.0%, p < 0.001).

Hospital stay was significantly longer in the Early group than the Immediate group (median 12.0 vs. 4.0 days; mean 13.2 ± 1.66 vs. 4.9 ± 0.10 days; both p < 0.001), and this remained significant after IPTW adjustment (p < 0.001). Compared to the Delayed group, the Early group had no significant difference in unadjusted median stay (12.0 vs. 13.0 days, p = 0.238), but after adjustment, stays were shorter in the Early group (median 10.0 vs. 12.0 days; mean 11.0 ± 0.50 vs. 15.0 ± 0.76 days; both p < 0.001). ICU stay did not differ significantly between Early and Immediate or between Early and Delayed groups after adjustment.

Discussion

Although LC has become the cornerstone treatment for AC, the ideal timing for surgery has remained a topic of controversy over the past decades and continues to be debated [3, 9, 21, 22]. While early surgery within 72 h of symptom onset has been proposed, this approach is not always feasible due to resource constraints or specific circumstances, such as surgical restrictions during control periods for special infectious diseases. Moreover, TG13 and TG18 recommend management strategies for AC based on disease severity. For patients with Grade II (moderate) AC who do not respond to initial medical treatment, and for those with Grade III (severe) AC—where organ support and stabilization of the general condition are critical—a combination of medical therapy and gallbladder drainage is recommended. In these cases, a planned cholecystectomy should be considered once the acute inflammation has resolved [22]. Furthermore, performing emergency surgery in high-risk patients can result in serious complications, further complicating treatment planning [23].

The optimal timing of LC following PTGBD remains unstandardized, with varying definitions and inconsistent results. Choi et al. reported that patients who underwent delayed LC five days after PTGBD had better outcomes than those who underwent LC within 72 h [15]. Similarly, Han et al. found that the perioperative complication rate was higher in patients who underwent LC within 72 h after PTGBD compared to those who had LC beyond 72 h [23]. In contrast, Jung et al. reported that scheduled LC following PTGBD (with a mean interval of seven days) did not reduce the rate of open conversion or postoperative hospital stay compared to emergency LC without PTGBD [24]. In 2019, Sakamoto et al. analyzed patients undergoing cholecystectomy following PTGBD for AC using a national database in Japan and concluded that performing cholecystectomy between 7 and 26 days after PTGBD may be reasonable in terms of mortality and morbidity [25]. The Tokyo Guidelines do not establish the optimal timing of LC after PTGBD due to ongoing controversy but emphasize that waiting 4 to 6 weeks after PTGBD may lead to better surgical outcomes [3, 9]. Traditionally, acute care surgeons at our institute follow the Tokyo Guidelines for managing AC. According to our flowchart for managing AC after PTGBD insertion, delayed LC is performed at least six weeks after drainage—often longer if the patient’s general condition is poor or if repeated attacks occur during the drainage period—rather than opting for early surgery. As a result, the proportion of patients in the Early group within our cohort is significantly lower (1.25%).

PTGBD is a key intervention for managing AC, particularly in high-risk patients who are unsuitable for immediate surgery. It provides a viable alternative to emergency cholecystectomy and serves as a bridge to definitive surgical treatment [3]. One of the primary benefits of PTGBD is its ability to provide rapid symptom relief, including pain, fever, and sepsis, by decompressing the inflamed gallbladder [25]. Additionally, PTGBD has been associated with lower immediate mortality and morbidity rates, making it a valuable option for critically ill patients with severe comorbidities. Moreover, for non-surgical candidates, PTGBD serves as a definitive management strategy, reducing the need for invasive procedures [26]. Despite its advantages, PTGBD is associated with several drawbacks. Catheter-related complications, including bile leakage, tube dislodgement or blockage, and infection, remain common concerns [27, 28]. In addition, prolonged dependence on PTGBD without definitive surgical intervention may result in recurrent episodes of cholecystitis, further increasing the overall healthcare burden [29, 30]. The financial impact of PTGBD should also be considered, as the procedure necessitates specialized interventional radiology expertise, additional hospital resources, and continuous follow-up care, all of which contribute to elevated healthcare costs [31]. In our study cohort, patients who underwent PTGBD were generally older, with a significantly higher proportion aged over 70 years (46.1% vs. 14.2%), and had a greater prevalence of multiple comorbidities (14.0% vs. 2.0%). However, only a limited number of these patients proceeded to delayed LC following PTGBD. For those deemed unsuitable for definitive surgical treatment, PTGBD likely served as a definitive management strategy due to their poor overall condition. Meanwhile, the remaining patients who underwent surgery were frequently treated with open cholecystectomy instead of minimally invasive surgery.

Since PTGBD can rapidly improve a patient’s overall condition, performing definitive surgery within a short period after PTGBD placement, as part of the same treatment course, may be a reasonable and feasible approach. This strategy could help minimize the decline in quality of life associated with prolonged PTGBD placement, reduce the risk of complications, and lessen the burden on long-term healthcare resources and costs. Additionally, delayed LC after PTGBD can present surgical challenges due to adhesions and gallbladder fibrosis, increasing the likelihood of conversion to open surgery [32]. Although prior studies have suggested that performing LC too early after PTGBD may lead to poorer outcomes [24, 32], some literature has proposed an optimal timing for surgery, such as between 7 and 26 days after PTGBD, as suggested by Sakamoto et al., or beyond 216 h, as recommended by Inoue et al. [25, 33]. However, the key factor in determining surgical suitability remains the patient’s overall stability and thorough preoperative preparation. In our study cohort, the median interval for early LC after PTGBD was 5.8 days. Outcome analyses adjusted by IPTW revealed no significant difference in the conversion to laparotomy rate among the three groups. However, both the Immediate group (direct LC) and the Delayed group (delayed LC after PTGBD) had a significantly higher incidence of major complications. In addition, the overall hospital stay was longer in the Delayed group. Based on our findings, early LC after PTGBD yields favorable outcomes, shortens the overall treatment period, and may help optimize the utilization of medical resources.

In the present study, we employed IPTW to address baseline imbalances and reduce confounding inherent in the retrospective design. A major advantage of IPTW is its ability to retain the full study sample, in contrast to propensity score matching (PSM), which may exclude unmatched cases and thereby further reduce statistical power—an important consideration when the sample size is already limited [19]. Even in smaller cohorts, IPTW has been shown to effectively adjust for measured confounders and achieve covariate balance between treatment groups, mitigating selection bias[20]. Simulation studies suggest that when the sample size exceeds approximately 60 subjects, IPTW can yield lower bias and mean squared error than PSM, particularly when matching would result in substantial loss of observations [19]. Nevertheless, in very small samples, extreme propensity scores may produce unstable estimates; therefore, we applied stabilized weights to reduce variance inflation and improve estimate precision [34].

This study has several limitations. First, its retrospective design introduces potential biases, such as selection bias and variations in patient management. Second, data were collected from a single center. Although this ensures that all patients were treated with a similar protocol and standard of care, it may limit the generalizability of the findings to other institutions with different surgical expertise or patient populations. Third, the sample size of the Early group was relatively small, which may affect the statistical power of the analysis. At our institute, delayed surgery after PTGBD has been the standard practice, even in cases where the patient’s general condition improved rapidly. Only a small number of patients underwent early surgery following PTGBD due to specific clinical indications. This approach contributed to the small sample size of the early surgery group. Additionally, despite using IPTW to adjust for confounding factors, some unmeasured variables may still influence the results. IPTW is also sensitive to model misspecification and extreme weights, which can increase variance and reduce reliability. In small sample sizes or highly imbalanced treatment groups, IPTW may lead to inaccurate estimates. Finally, the study did not evaluate long-term outcomes or patients’ quality of life (QoL), which are important factors in determining the best treatment approach especially those with prolong insertion of PTGBD. Most patients who underwent definitive cholecystectomy did not require long-term follow-up due to minimal postoperative impact on QoL. Although the delayed group was not the primary focus of this study, assessing their QoL is important for informed decision-making regarding early surgery. Future prospective, multicenter studies with larger sample sizes and extended follow-up periods are needed to validate these findings.

Conclusion

For patients requiring PTGBD, performing an early LC during the same admission may be a safe and effective alternative to prolonged delays in surgical intervention. Immediate LC remains the preferred treatment for eligible patients without prior PTGBD. However, in high-risk patients, excessive delays in LC following PTGBD may increase the risk of complications, extend hospitalization, and hinder recovery. Given the ongoing debate on the optimal timing of LC after PTGBD, our study explores the feasibility and safety of early surgery and demonstrates its potential benefits. This work provides a foundation for further prospective studies and randomized controlled trials with larger sample sizes to validate these findings and develop clear criteria for patient selection.

Author contributions

I-Ming Kuo contributed to the design of the study, manuscript preparation, editing of the manuscript, and is the primary writer of the manuscript. Sheng-Yu Chan contributed to data collection. Erh-Hao Liu, Jen-Fu Huang, Chih-Po Hsu, Chun-Hsiang Ou Yang, Shih-Ching Kang, Yu-Pao Hsu, and Chi-Hsun Hsieh contributed to the review and revision of the manuscript. Yi-Wen Hong contributed to the acquisition of data, and the coordination and correspondence of the study.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This retrospective analysis was approved by the Chang Gung Medical Foundation Institutional Review Board (201900056B0), which judged that written informed consent of the patients or their family was not necessary for this kind of retrospective study.

Competing interests

The authors declare no competing interests.