Introduction

Common bile duct stones (CBDSs) are present in approximately 10–20% of patients with symptomatic gallstones [1]. If left untreated, CBDSs can lead to complications in up to 25% of cases within four years, including acute pancreatitis, acute cholangitis, and biliary tract obstruction [2]. Accordingly, international guidelines recommend endoscopic retrograde cholangiopancreatography (ERCP) with stone extraction for all patients with CBDSs [1, 3, 4]. However, spontaneous passage of CBDSs through the papilla occurs in 6–27% of patients, indicating that some may undergo unnecessary ERCP [5–15]. Smaller stone size is a well-established predictor of spontaneous passage [5,8,9,11,12,14,15]. Nevertheless, findings regarding other potential predictors remain inconsistent, and most studies have been limited by small numbers of spontaneous passage events.

ERCP is an invasive procedure associated with a 6–15% risk of complications, including post-ERCP pancreatitis (PEP), bleeding, infection, perforation, and sedation-related adverse events, which can be potentially fatal [16,17]. Whether spontaneous passage of CBDSs influences the risk of post-ERCP complications has not been well studied.

To address these gaps, this study aimed to (1) identify clinical predictors of spontaneous passage of CBDSs and (2) evaluate the association between spontaneous passage of CBDSs and post-ERCP complications.

Materials and Methods

Results

A total of 404 patients diagnosed with CBDSs were included in this study. Spontaneous passage occurred in 113 patients (28%), confirmed by EUS in 16 patients and by ERCP in 97 patients (Fig 1). Among the 388 patients who underwent ERCP, the stone clearance rate was 93.8%. Of the study population, 50% were male. The mean age was 63 years, and 93% were symptomatic. The mean duration from diagnosis to procedure was 29 days. Approximately 80% of patients were diagnosed with CBDSs by CT (Table 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Study flow diagram.

Abbreviation: US, ultrasonography; CT, Computed tomography; MRI, Magnetic resonance imaging; MRCP, magnetic resonance cholangiopancreatography; EUS, endoscopic ultrasonography; CBDS, common bile duct stone; ERCP, endoscopic retrograde cholangiopancreatography.

https://doi.org/10.1371/journal.pone.0351242.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics.

https://doi.org/10.1371/journal.pone.0351242.t001

Predictive factors for spontaneous passage of CBDSs

In univariable analysis, younger age, presentation with acute pancreatitis, smaller CBD diameter, smaller CBDS size, single CBDS, distal CBDS location, lower ALP, higher AST, and higher ALT were associated with spontaneous passage of CBDSs. Lower pre-procedure ALP and normalization of pre-procedure ALP were also associated with spontaneous passage (Table 2). These findings remained consistent in a sensitivity analysis restricted to patients in whom the presence of CBDSs was confirmed by ERCP only and who did not undergo EUS prior to ERCP (S1 Table).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Univariable analysis of factors associated with spontaneous passage of common bile duct stones.

https://doi.org/10.1371/journal.pone.0351242.t002

The overall duration from diagnosis to procedures did not differ between the two groups. Subsequently, we performed an exploratory analysis to evaluate the risk of spontaneous CBDS passage across different time intervals, using patients who underwent ERCP within 7 days as the reference group. Among those who underwent ERCP within the first 28 days, the risk of spontaneous passage increased progressively with longer intervals, with the highest and statistically significant risk observed in patients who underwent ERCP 22–28 days after diagnosis (RR 1.89, 95% CI 1.13–3.15), as shown in S2 Table.

We also evaluated the potential impact of diagnostic modality on spontaneous passage and found that there was no significant difference in the prevalence of spontaneous passage of CBDSs across modalities (p = 0.74), and no significant association between diagnostic modality and spontaneous passage (Table S3).

In multivariable analysis, younger age, smaller CBDS size, and single CBDS remained significantly associated with spontaneous passage of CBDSs (Table 3). These associations remained consistent in a sensitivity analysis restricted to patients in whom the presence of CBDSs was confirmed by ERCP only and who did not undergo EUS prior to ERCP (S4 Table).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Multivariable analysis of factors associated with spontaneous passage of common bile duct stones.

https://doi.org/10.1371/journal.pone.0351242.t003

Association between spontaneous passage of CBDSs and post-ERCP complications

Overall post-ERCP complications tended to occur more frequently in patients with spontaneous passage of CBDSs compared with those without spontaneous passage (21.6% vs 13.4%, p = 0.052). PEP was significantly more common in patients with spontaneous passage than in those without (16.5% vs 7.6%, p = 0.01) (Table 4). However, the severity of PEP did not differ significantly between the two groups (mild, moderate, and severe = 87.5%, 12.5%, and 0% in the spontaneous passage group vs 86.4%, 9.1%, and 4.5% in the non-passage group; p = 0.66).

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Post-ERCP complications in patients with spontaneous passage of common bile duct stones compared with non-passage group.

https://doi.org/10.1371/journal.pone.0351242.t004

A multivariable analysis was performed to further evaluate the association between spontaneous passage of CBDSs and PEP. Confounding factors for adjustment were selected based on a directed acyclic graph constructed from our study findings and literature review [18,22]. The covariates included female sex, younger age, prior acute pancreatitis, asymptomatic presentation, CBD diameter, and normalization of bilirubin (Fig 2). After adjustment for these factors, spontaneous passage of CBDSs remained an independent factor associated with PEP, with a risk ratio of 2.34 (95% CI, 1.15–4.75).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Directed acyclic graph illustrating the causal pathways between spontaneous passage of common bile duct stones and post-ERCP pancreatitis.

Factors considered confounders are variables with arrows directed toward both spontaneous passage of CBDSs and post-ERCP pancreatitis. Abbreviations: ERCP, endoscopic retrograde cholangiopancreatography; CBDS, common bile duct stone; CBD, common bile duct.

https://doi.org/10.1371/journal.pone.0351242.g002

We further compared procedure-related factors between patients with and without PEP and found that pancreatic duct cannulation occurred more frequently among those who developed PEP (18.4% vs 3.4%, p < 0.01) (S5 Table). No pancreatic duct contrast injection was performed, and no rectal nonsteroidal anti-inflammatory drugs were used in this cohort. When procedural factors—including pancreatic duct cannulation, pancreatic duct stent placement, and precut sphincterotomy—were added to the previous multivariable model, spontaneous passage of CBDSs remained independently associated with PEP (RR 2.48, 95% CI 1.25–4.92).

Discussion

In this retrospective cohort study, which included a relatively large number of patients with spontaneous passage of CBDSs (113 patients, 28%), the prevalence was slightly higher than that reported in previous studies, which ranged from 6% to 27% [5–15], and also higher than the pooled prevalence of 15.7% (95% CI 11.5%–21.2%) reported in a recent meta-analysis [23]. The higher prevalence likely reflects the longer mean duration from diagnosis of CBDSs to procedures (29 days), compared with 4–24 days in prior studies [7,9,11,12,14,15], which could allow more time for CBDSs to pass through the papilla spontaneously. The relatively prolonged interval in our cohort reflects both referral patterns to our tertiary center—where patients are often transferred after stabilization of the initial acute phase —and limited ERCP and EUS capacity, as these procedures are available in only a few centers in our region, thereby contributing to longer waiting times. This unintentional delay provided a unique window to observe the natural history of stone passage.

We identified several predictors associated with spontaneous passage of CBDSs. Smaller CBDS size was confirmed as a strong predictor of spontaneous passage, consistent with prior reports [5,8,9,11,12,14,15]. A recent meta-analysis similarly demonstrated this association, reporting a mean CBDS size of 3.8 mm among patients with spontaneous passage [23]. In our cohort, the mean CBDS size in the spontaneous passage group was slightly larger than that reported in the meta-analysis (4.9 mm). The appropriate cut-off for CBDS size used to predict spontaneous passage has varied across studies. Previous investigations have proposed thresholds ranging from 2 to 8 mm, each with differing performance [5–9,11,12,14,15]. Our study provides additional insight by demonstrating that each 1-mm increase in CBDS size reduces the likelihood of spontaneous passage, with a risk ratio of 0.78. This continuous effect may be incorporated into future clinical prediction models for more precise risk estimation. We also found that having a single CBDS was associated with spontaneous passage, consistent with previous reports [8,12,23]. Another imaging characteristic identified in prior studies is CBD diameter, which tends to be smaller in patients with spontaneous passage [12,23]. This was observed in our univariable analysis; however, the association was not retained in the multivariable model.

Previous studies have reported that patients with spontaneous passage tend to be younger [13,23], and our multivariable analysis confirmed this association. Increasing age was significantly associated with a lower likelihood of spontaneous passage, with the risk decreasing by an RR of 0.88 for every 10-year increase. The underlying mechanism of this association remains unclear and requires further investigation. Another demographic factor previously studied is female sex, which was reported in one study to be associated with a lower likelihood of spontaneous passage of CBDSs[11]. However, our findings, consistent with the meta-analysis, showed that sex was not a significant predictor [23].

Presentation with acute pancreatitis was not identified as a significant predictor in a recent meta-analysis (RR 1.60, 95% CI 0.81–3.17) [23]. However, a multicenter retrospective study reported that acute pancreatitis was a significant predictor of spontaneous passage in multivariable analysis (OR 2.78, 95% CI 1.14–6.73) [11]. In our study, univariable analysis similarly showed that acute pancreatitis increased the likelihood of spontaneous passage (RR 1.64, 95% CI 1.15–2.34). Nonetheless, in our more conservative multivariable model that adjusted for all potential predictive factors, acute pancreatitis was not a significant predictor (RR 1.28, 95% CI 0.87–1.87). For another clinical manifestation, acute cholangitis, our finding was consistent with the meta-analysis, which showed that acute cholangitis was not a significant predictor of spontaneous passage [23].

Two studies have reported that a longer interval from diagnosis to ERCP is associated with a higher likelihood of spontaneous passage of CBDSs in multivariable analysis [11,12], and this association was also consistent with a recent meta-analysis [23]. In contrast, our study did not observe this relationship in either univariable or multivariable analysis. However, our exploratory analysis demonstrated that, among patients who underwent ERCP within 28 days of diagnosis, the risk of spontaneous CBDSs passage increased progressively with longer intervals, with the highest and statistically significant risk observed in those who underwent ERCP 22–28 days after diagnosis. This may explain why prior studies—with mean procedure intervals of 5–6 days—identified duration as a significant predictor [11,12], whereas our study, which had a substantially longer mean interval of 29 days, did not show this association in the overall analysis.

Most studies have focused on predictive factors present at the time of CBDSs diagnosis; however, the likelihood of spontaneous passage of CBDSs may change over time, particularly when the interval between diagnosis and the procedure is prolonged. Because several parameters can change dynamically, it is also important to explore potential predictors measured closer to the procedure. A retrospective study from another center in Thailand demonstrated that resolution of clinical symptoms (absence of fever and abdominal pain) in patients with CBDSs and cholangitis predicted spontaneous passage [8]. A study from Portugal found that lower pre-ERCP total bilirubin increased the likelihood of spontaneous passage and proposed a cut-off of ≤ 2 mg/dL as a clinical decision guide [15]. In contrast, our study showed that pre-procedure bilirubin was not a significant predictor in either univariable or multivariable analyses. This discrepancy may be explained by the substantially longer mean interval between diagnosis and procedures in our cohort. Given the natural course of acute biliary obstruction, AST, ALT, and bilirubin levels typically decline spontaneously within a few days [24]; thus, in our cohort—with a relatively prolonged interval—the median bilirubin levels were already within the normal range and did not differ between groups. Furthermore, caution should be taken when using pre-procedure bilirubin as a predictor, as bilirubin levels may be affected in patients with bilirubin metabolism disorders or inherited hemoglobinopathies such as thalassemia, in which gallstone-related diseases are relatively common [25]. Our study demonstrated that lower pre-procedure ALP was associated with spontaneous passage in univariable analysis, potentially reflecting the slower decline of ALP after biliary obstruction. However, this association was not significant after multivariable adjustment.

Previous studies have reported no difference in post-ERCP complications between patients with spontaneous passage and those without; however, the number of complication cases in the spontaneous passage group was relatively small, which may have limited the statistical power to detect a difference [8,12,14,26]. A recent retrospective cohort study reported complication rates of 8% versus 4% in patients with and without spontaneous passage, respectively, although statistical comparison was not provided [15]. In our study, the overall complication rate also tended to be higher in patients with spontaneous passage, but the difference was not statistically significant. Nonetheless, the prevalence of PEP was significantly higher in the spontaneous passage group. Most cases were mild and the severity distribution did not differ between groups. We subsequently performed a multivariable analysis to evaluate the association between spontaneous passage of CBDSs and PEP. After adjusting for relevant confounders identified using a direct acyclic graph, spontaneous passage remained significantly associated with PEP. The association persisted after further adjustment for procedure-related factors. This association may be related to mechanisms similar to that proposed for acute biliary pancreatitis, in which transient ampullary obstruction from passing stones triggers pancreatic inflammation [27], followed by a second hit from mechanical manipulation during ERCP. Further studies are needed to clarify the mechanisms underlying this observation.

Our study highlights not only the issue of unnecessary ERCP but also the increased risk of PEP in patients with spontaneous passage of CBDSs. Developing and validating risk prediction tools that incorporate predictors identified in our study and others may help clinicians more accurately assess an individual patient’s likelihood of spontaneous passage and avoid the potential risks associated with ERCP. For patients with a high predicted likelihood of spontaneous passage, a highly accurate noninvasive test—such as EUS or MRCP—could be considered before proceeding to ERCP. These modalities show comparable diagnostic performance for detecting CBDSs, with EUS demonstrating a sensitivity of 95% and specificity of 97%, and MRCP demonstrating a sensitivity of 93% and specificity of 96% [28]. Supporting this, all 9 patients in our cohort who underwent EUS shortly before ERCP (on the same day or within one day) demonstrated persistent CBDSs on EUS, which were subsequently confirmed at ERCP. In clinical practice, EUS may offer a practical advantage, as it can be performed immediately before ERCP in the same procedural session, allowing real-time reassessment of ductal clearance and facilitating avoidance of unnecessary ERCP. In cases where noninvasive testing confirms the absence of stones, expectant management without ERCP is likely safe. Supporting this approach, a prospective study of patients with clinically suspected CBDSs but negative MRCP findings, followed for a mean of 10 months, reported only one case of recurrent CBDSs among 50 patients; this occurred in an individual with gallstones who had not undergone cholecystectomy [29]. However, in patients with CBDSs and acute cholangitis, the use of repeat noninvasive confirmation or expectant management—even when spontaneous passage is suspected—should be carefully weighed against potential risks and benefits, and further data on safety outcomes in this population are warranted. At present, ERCP remains the standard treatment recommended by international guidelines, particularly in patients with moderate to severe cholangitis requiring urgent intervention [30].

To our knowledge, this is the first study to demonstrate an increased risk of PEP in patients with spontaneous passage of CBDSs. Our cohort included a relatively large number of patients with spontaneous passage, and we applied robust statistical methods to evaluate both the predictors of spontaneous passage and its independent association with PEP. However, several limitations should be acknowledged. First, this was a retrospective study. Second, it was conducted at a single center, and the relatively long interval from diagnosis to procedures in our cohort may limit generalizability of the findings to other settings, particularly where procedures can be performed more rapidly. Third, CBDS confirmation was performed using heterogeneous methods—either EUS or ERCP. Nevertheless, EUS has high sensitivity and specificity for diagnosing CBDSs [28], making it highly comparable to ERCP. In addition, sensitivity analyses restricted to patients whose CBDSs were confirmed by ERCP alone yielded consistent results, supporting the robustness of our findings. Fourth, the diagnosis of CBDSs in our cohort relied on multiple imaging modalities. However, our exploratory analysis showed that diagnostic modality did not significantly influence the likelihood of spontaneous CBDSs passage. Another consideration is that each imaging modality carries a different potential false-positive rate, which may confound the observed prevalence of spontaneous CBDS passage. A study in which EUS was performed within 3 days prior to ERCP reported a false-positive rate of 1.6% (2 of 126 positive cases) [31]. Another study evaluating transabdominal ultrasonography and MRCP, performed within 4 hours to 2 weeks (mean 18 hours) before direct cholangiography, found that MRCP had a false-positive rate of 9.7% (3 of 31 cases), mainly due to misinterpretation of a prominent ampullary sphincter as a distal bile duct stone, whereas ultrasonography showed no false positive cases (0 of 13 cases) [32]. A study evaluating CT, with a mean interval of 7 days from CT to ERCP (range: same day to 15 days), reported false-positive rates ranging from 16.7% to 18.2% for non-contrast CT and from 27.8% to 35.0% for contrast-enhanced CT. These false-positive findings were attributed to opacified blood vessels adjacent to the bile duct or contrast enhancement of the bile duct mucosa [33]. However, because these studies included varied time intervals between imaging and ERCP, it is difficult to distinguish true false-positive findings from spontaneous stone passage. Notably, a study evaluating non-contrast CT performed within 1 hour prior to ERCP reported no false-positive cases among 15 patients with positive CT findings [34]. This suggests that the true false-positive rate of these imaging modalities may be lower than reported when diagnostic imaging is performed shortly before ERCP. Lastly, intraductal ultrasound or cholangioscopy was not performed during ERCP, so the possibility of missed small residual stones cannot be entirely excluded.

In conclusion, this study demonstrated that younger age, smaller CBDS size, and a single CBDS are significant predictors of spontaneous passage of CBDSs. Integrating these factors into clinical decision-making may help identify patients with a high likelihood of spontaneous passage who would benefit from noninvasive evaluation with EUS or MRCP to confirm the persistence of CBDSs. This approach has the potential to avoid ERCP in up to one-quarter of patients. Such a strategy would not only reduce unnecessary ERCP procedures but also decrease the risk of PEP, which we found to be associated with spontaneous passage of CBDSs. Future studies are warranted to develop and externally validate clinical prediction models for spontaneous passage of CBDSs, incorporating these factors to better guide individualized decision-making.

Supporting information