Abstract
Purpose
To identify clinical and radiological prognostic factors associated with delayed surgery after diagnosis of a Crohn’s disease-related intra-abdominal abscess.
Methods
We conducted a retrospective single-center study including patients with Crohn’s disease and confirmed intra-abdominal abscess, managed between 2012 and 2022. The primary outcome was delayed surgery related to the index abscess, defined as bowel resection or surgical drainage performed between day 15 and one year after diagnosis. Clinical, biological, therapeutic, and imaging data were collected, and initial imaging studies were systematically reviewed. Hazard ratios (HR) were estimated using univariate and multivariate Cox models. Kaplan-Meier survival analyses and log-rank tests assessed time to surgery.
Results
Among 145 patients included, 59 underwent surgery within one year. In multivariate analysis, larger abscess long-axis diameter (HR = 1.22; 95% CI [1.09–1.35]; p < 0.001) and older age at abscess diagnosis (HR = 1.15; 95% CI [1.01–1.31]; p = 0.029) were associated with higher surgical risk. In contrast, older age at Crohn’s disease diagnosis (HR = 0.83; 95% CI [0.72–0.96]; p = 0.011), collection morphology with mass effect compared with infiltrative morphology (HR = 0.80; 95% CI [0.65–0.99]; p = 0.038), and intravenous (IV) antibiotic therapy (HR = 0.79; 95% CI [0.64–0.97]; p = 0.024) were associated with reduced surgical risk. None of these factors significantly influenced time to surgery in Kaplan–Meier survival curve analysis.
Conclusion
Larger abscesses and older age increased surgical risk, while IV antibiotics, well-defined morphology, and later Crohn’s diagnosis reduced it.
Keywords: Crohn disease, abscess, surgery, Magnetic Resonance Imaging, computed tomography
Graphical Abstract
Highlights
-
•
Clinical and radiologic predictors of surgical outcomes after Crohn’s abscess.
-
•
Age, abscess morphology, antibiotics, and Crohn’s diagnosis timing influence surgical risk.
-
•
Identified predictors help radiologists guide timely medical versus surgical management.
1. Introduction
Intra-abdominal abscesses are a recognized and potentially serious complication of fistulizing Crohn’s disease, arising from transmural inflammation [1]. Management typically involves targeted antibiotic therapy against enteric bacteria [2], often combined with drainage via percutaneous, endoscopic, or surgical approaches [3], [4]. However, there is limited consensus on the optimal timing and method of intervention [5].
Cross-sectional imaging techniques, including computed tomography (CT) and magnetic resonance imaging (MRI), play a key role in evaluating abscess location, size, and complexity, thereby guiding decisions on whether percutaneous drainage might be appropriate [6], [7]. These modalities also help identify extra-enteric complications that are not visible on endoscopy [8]. Accurate analysis of imaging findings can facilitate earlier recognition of cases where surgery may be warranted, potentially reducing the risk of perforation or sepsis [9].
Some studies report similar rates of complications and recurrence regardless of strategy [10]. However, while surgical resection of the perforated bowel segment and associated fistulous tracts provides a more definitive resolution of abscesses, an emergency surgery may require extensive ileal resection and a temporary stoma, significantly impacting the patient’s quality of life [11]. Furthermore, immunomodulatory and biologic therapies must often be reintroduced to control disease activity [12].
Although size-based thresholds have been proposed to distinguish between medical and surgical management [13], they need to be further validated, and additional criteria should be explored to guide a more personalized treatment. The objective of this study was to identify both clinical and radiological factors associated with delayed surgical management, defined as surgery performed between 15 days and one year after diagnosis, in patients with Crohn’s disease presenting with intra-abdominal abscesses.
2. Material and Methods
2.1. Inclusion and exclusion criteria
A retrospective observational study was conducted on patients hospitalized between 01/01/2012 and 01/01/2022 for intra-abdominal abscess associated with Crohn’s disease at Saint-Antoine Hospital, Paris. The cohort end date corresponded to the predefined data lock and was chosen to ensure complete one-year follow-up and feasibility of individual chart and imaging review. This retrospective study was approved by the Comité d′Éthique pour la Recherche en Imagerie Médicale (CERIM) of the Collège des Enseignants en Radiologie de France (CERF), under IRB number CRM−2404–404. Patients were identified based on hospitalization records using ICD−10 codes for Crohn's disease (K50.x) and intra-abdominal abscess (K63.0). Each case was then individually reviewed, including review of medical records and available imaging, to confirm eligibility.
Inclusion criteria included age ≥ 18 years, a confirmed diagnosis of Crohn's disease (CD), available imaging data, and patient management at Saint-Antoine Hospital.
Exclusion criteria comprised abscesses occurring less than 3 months after any bowel surgery, perianal abscesses, peritonitis, no imaging available, no inflammatory bowel disease (IBD) on review of medical records and no abscesses. The flow chart outlining the selection process is presented in Fig. 1.
Fig. 1.
Patient selection flowchart and delayed-surgery time window. Percentages are calculated using the preceding analysis denominator unless otherwise stated.
The primary outcome was delayed abdominal surgery related to Crohn’s disease and the index intra-abdominal abscess, defined as any abdominal surgical intervention, including bowel resection or surgical abscess drainage, performed between 15 days and one year after abscess diagnosis. The aim of the study was to identify the clinical, biological, therapeutic, and radiological factors associated with this primary outcome. Time to surgery was assessed as an additional exploratory analysis.
2.2. Imaging analysis
For each patient, an initial imaging examination (MRI or CT) was identified corresponding to the examination performed at the start of management for an intestinal abscess. The database consists of CT scans (78%; 113/145), most often including at least portal-phase acquisition (98%; 111/113) or, if unavailable, performed without contrast injection. MRI scans represent 22% (32/145), each containing at least axial and coronal T2 sequences without fat saturation, a dynamic coronal sequence, an axial diffusion sequence, and frequently two-plane (axial and coronal) T1 post-contrast sequences (78%; 25/32). These imaging examinations were reviewed by a junior radiologist (*blinded*) and a senior radiologist (*blinded*), both blinded to clinical outcomes and biological data. When annotations differed, two senior radiologists (> 5 years of experience in abdominal imaging) conducted a joint review (*blinded*).
The criteria evaluated on initial examination were as follows: abscess dimensions (craniocaudal, anteroposterior, and transverse diameters in mm), presence of an enhancing abscess wall after contrast administration, infiltrative nature or mass effect of abscess defined by its shape in space and its relationship with intra-abdominal anatomical structures (Fig. 2), presence of accessory collections, air within the abscess, presence of a predominantly inflammatory component (liquid <50%), type of fistula (simple: blind/sinus, entero-enteric, or entero-colic; complex: ≥2 tracts), presence of stenosis (defined as an upstream dilatation >3 cm) and length of inflammatory bowel involvement (in cm). An abscess with a mass effect typically appears rounded and displaces adjacent structures, whereas an infiltrative abscess does not exert mass effect but instead insinuates itself between surrounding tissues. In MRI examinations, air within the abscess was recorded as absent unless it was clearly visible.
Fig. 2.
Abscess Examples: Well-Defined vs. Infiltrative. The first row shows an abscess (star) with well-defined morphology and pathological ileal bowel (arrow). The second row displays an infiltrative abscess with multiple poorly defined loculations (star).
2.3. Clinical and biological data
Clinical and biological variables were collected from the SUVIMIC registry (a prospective clinical database of all patients with IBD evaluated by Saint Antoine Hospital digestive disease medical staff) [14], [15]. Additional data were manually extracted from files when necessary to ensure completeness of the dataset. For each selected patient, clinical and biological data at initial examination were collected. These data included age, BMI, smoking status, age at diagnosis of Crohn's disease, duration between diagnosis and occurrence of abscesses, presence and type of perianal lesions, phenotype and location of disease according to the Montreal classification [16], presence of associated primary sclerosing cholangitis, as well as hemoglobin, C-reactive protein (CRP) and albumin levels. The Montreal behavior variable reported in the descriptive table corresponded to the phenotype recorded in the clinical registry before or at the time of the index episode and was analyzed without incorporating the current abscess episode itself. Systemic corticosteroid and immunomodulator exposure at the time of abscess diagnosis, and the number of previous biologic lines, were also recorded. When available, anti-TNF exposure after abscess management was specifically collected. Finally, antibiotic therapy (intravenous or oral) and the use of enteral and/or parenteral nutrition during abscess management were documented. Intravenous antibiotic therapy was analyzed as a separate variable.
Initial medical management was not governed by a strict written protocol during the entire study period. Decisions regarding antibiotics, nutritional support, percutaneous drainage, surgical drainage, and timing of surgery were made case by case, usually after discussion among gastroenterologists, digestive surgeons, radiologists, and interventional radiologists. Percutaneous drainage was considered according to abscess size, morphology, accessibility, and clinical status. This real-life management strategy may introduce treatment-selection bias and was therefore explicitly considered when interpreting drainage-related analyses.
2.4. Statistical analysis
Statistical methods were applied using Python libraries, including lifelines for survival analyses, to assess the impact of various clinical, radiological, and biological variables. Continuous variables, expressed as median (interquartile range 25–75), were analyzed using the Mann-Whitney, Wilcoxon test. Qualitative variables, expressed as number (percentage), were compared using the Chi-squared test with Yates correction. For each variable included in the univariate Cox model, the Hazard Ratio (HR) and its 95% confidence intervals (95% CI) were calculated. A p-value of less than 0.05 was considered statistically significant. Inter-individual variability was assessed using a kappa concordance test and the intra-class correlation coefficient.
For the multivariate analysis, a Cox proportional hazards model was used. Variable selection was based on bootstrap inclusion frequency analysis, as the event-to-variable ratio was moderate (<10). We did not restrict variable selection solely to variables significant in univariate analysis, because univariate screening may exclude clinically relevant predictors and does not account for confounding. Variables with an inclusion frequency greater than 50% were retained in the Cox model (Supplementary Table 1) [17]. Continuous variables, including abscess long axis, were analyzed as continuous predictors in Cox models. Kaplan-Meier survival curves were generated, and the log-rank test was used to compare survival outcomes between groups. Two abscess long-axis thresholds of 30 mm and 50 mm were used for Kaplan-Meier visualization only.
3. Results
3.1. Clinical characteristics
Based on discharge diagnosis codes, we initially identified 222 Crohn's disease patients hospitalized for intra-abdominal abscesses from 2012 to 2022. A total of 145 patients were included and divided into two groups according to their operative status. The first group, comprising 59 patients (41%), underwent surgery within one year of intra-abdominal abscess diagnosis, while the second group, comprising 86 patients (59%), did not. The flow chart outlining the selection process is presented in Fig. 1. The clinical, biological, radiological and Crohn's disease history characteristics of the operated/non-operated groups are summarized in Table 1. No significant differences were observed between groups in terms raof age, BMI, smoking status, or age at Crohn’s disease diagnosis. Similarly, time from diagnosis to abscess onset did not differ significantly. Disease phenotype and location, based on the Montreal classification, were comparable between groups. The Montreal behavior variable used in the analysis reflected the historical phenotype recorded before or at the time of the index episode and did not incorporate the current abscess episode itself. While a slightly higher proportion of patients in the surgery group had a stricturing phenotype (B2: 19% vs. 8%), this was not statistically significant (p = 0.9). Prior history of bowel or perianal surgery was also not significantly associated with the risk of delayed surgery.
Table 1.
Clinical, biological, radiological and therapeutic features.
| Delayed surgery (day 15–12 months) | ||||||
|---|---|---|---|---|---|---|
| Patients | All (n = 145) | Yes (n = 59) | No (n = 86) | p | ||
| Clinical parameters | ||||||
| Age (years) | 35 [30–44] | 37 [31–45] | 34 [29–41] | 0.23 | ||
| BMI (kg/m²) | 20.8 [18.3–23.8] | 21.3 [17.7–23.9] | 19.9 [18.3–23.8] | 0.3 | ||
| Smoking | No | 71 (49) | 28 (47) | 43 (50) | 0.43 | |
| Active | 35 (24) | 12 (20) | 23 (27) | |||
| Former | 39 (27) | 19 (32) | 20 (23) | |||
| Crohn's disease | ||||||
| Age at Crohn's diagnosis (years) | 21 [17–26] | 21 [16–25] | 22 [17–26] | 0.36 | ||
| Time from diagnosis to abscess onset (years) | 7 [1–15] | 10 [3–17] | 6 [1–12] | 0.2 | ||
| Perianal lesions | Primary | 3 (2) | 2 (3) | 1 (1) | 0.09 | |
| Secondary | 42 (29) | 20 (34) | 22 (26) | |||
| Not specified | 9 (6) | 4 (7) | 5 (6) | |||
| Montreal classification: Location | L1 | 58 (40) | 24 (41) | 34 (40) | 0.56 | |
| L2 | 5 (3) | 1 (2) | 4 (5) | |||
| L3 | 75 (52) | 33 (56) | 42 (49) | |||
| Upper gastrointestinal involvement (Montreal L4) | 23 (16) | 12 (20) | 11 (13) | |||
| Historical Montreal behavior before/index abscess episode | B1 | 19 (13) | 8 (13) | 11 (13) | 0.9 | |
| B2 | 18 (12) | 11 (19) | 7 (8) | |||
| B3 | 67 (46) | 32 (54) | 35 (41) | |||
| Primary sclerosing cholangitis | 1 (100) | 0 | 1 (100) | 1 | ||
| History of bowel resection | 31 (21) | 16 (27) | 15 (17) | 0.23 | ||
| History of perianal surgery | 24 (17) | 8 (14) | 16 (19) | 0.56 | ||
| Biology | ||||||
| Hb (g/dL) | 12 [10.8–13] | 11.4 [10.1–12.7] | 12.2 [10.9–13.1] | 0.2 | ||
| CRP (mg/L) | 102 [60−160] | 110 [72−171] | 94 [54−142] | 0.13 | ||
| Albuminemia (g/L) | 31 [28–35] | 30 [27–33] | 31 [28–36] | 0.07 | ||
| Imaging | ||||||
| Abscess size (ellipsoid formula, mm^3) | 103 [44−321] | 103 [48−410] | 103 [38−217] | 0.8 | ||
| Abscess long axis (maximum diameter, mm) | 36 [30–44] | 38 [29–54] | 35.5 [23.5–47] | 0.21 | ||
| Enhancing abscess wall | 121 (83) | 48 (81) | 73 (85) | 0.74 | ||
| Presence of accessory collections | 63 (43) | 31 (53) | 32 (37) | 0.10 | ||
| Collection with mass effect | 67 (46) | 23 (39) | 44 (49) | 0.20 | ||
| Air within the collection | 74 (51) | 27 (46) | 47 (55) | 1.00 | ||
| Predominant fibro-inflammatory component | 46 (32) | 19 (32) | 27 (31) | 1.00 | ||
| Fistula type | Simple | 98 (68) | 37 (63) | 61 (71) | 0.39 | |
| Complex | 47 (32) | 12 (20) | 23 (27) | |||
| Fistula type | Entero-enteric | 27 (19) | 9 (15) | 18 (21) | 0.41 | |
| entero-colic | 15 (10) | 6 (10) | 9 (10) | |||
| Sinus tract | 103 (71) | 44 (75) | 59 (69) | |||
| Presence of stenosis | 39 (27) | 15 (25) | 24 (28) | 0.89 | ||
| Treatment | ||||||
| Immunosuppressants (thiopurines, methotrexate) | 29 (20) | 13 (22) | 16 (19) | 0.77 | ||
| Number of biotherapy lines received | 0 | 63 (44) | 24 (40) | 39 (45) | 0.89 | |
| 1 | 45 (31) | 18 (31) | 27 (31) | |||
| 2 | 18 (12) | 8 (14) | 10 (12) | |||
| > =3 | 19 (13) | 9 (15) | 10 (12) | |||
| Systemic corticosteroids at abscess diagnosis | 26 (18) | 13 (22) | 13 (15) | 0.5 | ||
| Intravenous antibiotic therapy | 102 (70) | 38 (64) | 64 (74) | 0.22 | ||
Continuous variables, expressed as median (interquartile range 25–75), were analyzed using the Mann-Whitney, Wilcoxon test. Qualitative variables, expressed as number (percentage), were compared using the Chi-squared test with Yates correction.
3.2. Biological characteristics
No significant differences were identified between groups regarding hemoglobin, C-reactive protein (CRP), and albumin levels at the initial examination (Table 1).
3.3. Imaging characteristics
The characteristics analyzed on each patient's initial images are summarized in Table 1. There was no significant difference in the proportion of all the different characteristics between the groups who had undergone surgery during the year and those who had not: We observed no statistically significant difference between groups notably concerning the maximum abscess diameter (p = 0.21) and the presence of a mass effect by the collection (39% vs. 49%; p = 0.2). Although no difference in abscess size was detected by the Mann-Whitney test (p = 0.21).
3.4. Inter-individual variability of imaging characteristics
The results of the kappa agreement test are presented in Table 3. Agreement was perfect or strong for all characteristics, except for the presence of enhancing abscess wall (κ = 0.57) and for the classification of fistula types, whether simple/complex or inter-loop/entero-colic/blind (κ = 0.53 for both) which showed moderate agreement. These results highlight areas of reliable reproducibility, particularly for collections with a mass effect or infiltrative nature, while also pointing to moderate agreement in more complex classifications such as fistula type.
Table 3.
Inter-individual agreement.
| Enhancing abscess wall | 0.57 |
|---|---|
| Presence of accessory collections | 0.66 |
| Collection with mass effect | 0.68 |
| Air within the collection | 1 |
| Predominant fibro-inflammatory component | 0.74 |
| Fistula type: complex/simple | 0.53 |
| Fistula type: entero-enteric/entero-colic/blind | 0.53 |
| Presence of stenosis | 0.80 |
3.5. Treatment of Crohn's disease and intra-abdominal abscesses
We found no statistically significant differences between groups regarding immunosuppressant exposure (13/59 [22%] in the operated group vs 16/86 [19%] in the non-operated group; p = 0.77) or systemic corticosteroid exposure at abscess diagnosis (13/59 [22%] vs 13/86 [15%]; p = 0.5), as well as the administration of intravenous antibiotic therapy (38/59 [64%] vs 64/86 [74%]; p = 0.22). Percutaneous drainage was performed in 15/145 patients (10%), including 9/59 (15%) in the delayed-surgery group and 6/86 (7%) in the non-surgery group (p = 0.18). Among patients who remained surgery-free at one year, follow-up imaging around day 40 was documented in 79/86 (92%). Complete abscess resolution was observed in 23/86 patients, resolution with persistence of an inflammatory mass in 31/86, and persistent/residual abscess in 25/86. Anti-TNF follow-up exposure was documented in 117 patients. Overall, 85/117 (73%) received anti-TNF therapy during follow-up, including 38/55 (69%) in the delayed-surgery group and 47/62 (76%) in the non-surgery group (p = 0.54). These data were reported descriptively and were not included in the predictive model because they represented post-index therapeutic exposure.
3.6. Cox regression analysis
We used Cox proportional hazards models as univariate survival analyses to assess associations between radiological and biological variables and the likelihood of delayed surgery between day 15 and one year after abscess diagnosis. The abscess long axis, analyzed as a continuous variable, was identified as a significant predictor of surgical risk (HR = 1.48, 95% CI: 1.15–1.90, p = 0.002), highlighting the importance of abscess size in surgical outcomes.
In contrast, other variables, including the presence of accessory collections, type of fistula, enhancing abscess wall, air within the collection, fibroinflammatory components, stenosis, intravenous antibiotic therapy, age, CRP, albumin, age at Crohn’s disease diagnosis, and smoking, did not show statistically significant associations (Table 2, Fig. 3).
Table 2.
Results of Univariate and Multivariate Survival Analyses.
| Univariate Cox survival analysis |
Multivariate Cox survival analysis |
|||
|---|---|---|---|---|
| Variable | HR [95% CI] | p-value | HR [95% CI] | p-value |
| Clinical | ||||
| Sex | 1.12 [0.67; 1.88] | 0.672 | 1.04 [0.86; 1.25] | 0.698 |
| Age | 1.06 [0.83; 1.33] | 0.653 | 1.15 [1.01; 1.31] | 0.029 |
| BMI | 0.96 [0.74; 1.25] | 0.770 | 1.00 [0.90; 1.10] | 0.934 |
| Smoking | 0.77 [0.41; 1.46] | 0.431 | 0.95 [0.75; 1.20] | 0.658 |
| Crohn’s disease | ||||
| Age at IBD diagnosis | 0.82 [0.61; 1.11] | 0.203 | 0.83 [0.72; 0.96] | 0.011 |
| Biology | ||||
| CRP | 1.13 [0.90; 1.42] | 0.289 | 1.01 [0.91; 1.11] | 0.918 |
| Albumin | 0.79 [0.62; 1.02] | 0.070 | 0.94 [0.84; 1.06] | 0.328 |
| Imaging | ||||
| Enhancing abscess wall | 0.83 [0.43; 1.60] | 0.583 | 1.07 [0.84; 1.36] | 0.575 |
| Presence of accessory collection | 1.47 [0.88; 2.45] | 0.142 | 1.06 [0.86; 1.30] | 0.608 |
| Collection with mass effect or infiltrative | 0.71 [0.42; 1.19] | 0.191 | 0.80 [0.65; 0.99] | 0.038 |
| Air within the collection | 0.96 [0.57; 1.60] | 0.867 | 0.89 [0.72; 1.10] | 0.289 |
| Predominant fibroinflammatory component | 1.07 [0.62; 1.84] | 0.815 | 1.00 [0.81; 1.24] | 0.983 |
| Fistula type: simple or complex | 0.77 [0.46; 1.31] | 0.338 | 1.02 [0.78; 1.32] | 0.904 |
| Fistula type: entero-enteric, entero-colic, others | 1.21 [0.72; 2.03] | 0.481 | 1.02 [0.81; 1.29] | 0.879 |
| Stenosis | 0.92 [0.51; 1.64] | 0.767 | 0.87 [0.69; 1.10] | 0.234 |
| Abscess long axis (mm, continuous) | 1.48 [1.15; 1.90] | 0.002 | 1.22 [1.09; 1.35] | < 0.001 |
| Treatment | ||||
| Systemic corticosteroids | 0.95 [0.51; 1.75] | 0.863 | 0.98 [0.79; 1.23] | 0.893 |
| Immunosuppressants (thiopurines or methotrexate) | 1.16 [0.63; 2.15] | 0.634 | 1.17 [0.93; 1.49] | 0.180 |
| Number of biologic lines | 1.07 [0.84; 1.37] | 0.585 | 0.97 [0.87; 1.08] | 0.574 |
| Intravenous antibiotic therapy | 0.66 [0.39; 1.12] | 0.122 | 0.79 [0.64; 0.97] | 0.024 |
| Hemoglobin | 1.07 [0.83; 1.39] | 0.588 | 1.02 [0.93; 1.11] | 0.686 |
Fig. 3.
Univariable Cox regression analysis for delayed surgery. Forest plot showing hazard ratios (HRs), 95% confidence intervals (CIs), and p-values from univariable Cox proportional hazards models assessing variables associated with delayed surgery between day 15 and one year after abscess diagnosis. P-values correspond to Wald tests from the Cox regression models. Colored rows indicate variable domains, and filled markers indicate p < 0.05.
Multivariate analysis confirmed that abscess long axis remained significantly associated with an increased risk of surgery (HR = 1.22; 95% CI [1.09; 1.35]; p < 0.001). Older age was also associated with increased surgical risk (HR = 1.15; 95% CI [1.01; 1.31]; p = 0.029), whereas older age at IBD diagnosis was associated with reduced surgical risk (HR = 0.83; 95% CI [0.72; 0.96]; p = 0.011) (Table 2, Fig. 4). Collection morphology with mass effect, compared with infiltrative morphology, was associated with reduced surgical risk (HR = 0.80; 95% CI [0.65; 0.99]; p = 0.038), as was intravenous antibiotic therapy (HR = 0.79; 95% CI [0.64; 0.97]; p = 0.024). Other variables, including accessory collections, enhancing abscess wall, air within the abscess, fibroinflammatory components, fistula type, stenosis, and CRP, did not demonstrate statistically significant associations with surgical outcomes.
Fig. 4.
Multivariable Cox regression analysis for delayed surgery. Forest plot showing adjusted hazard ratios (HRs), 95% confidence intervals (CIs), and p-values from the multivariable Cox proportional hazards model for delayed surgery between day 15 and one year after abscess diagnosis. Variables were retained according to bootstrap inclusion frequency as described in the Methods. P-values correspond to Wald tests from the Cox regression model. Colored rows indicate variable domains, and filled markers indicate p < 0.05.
3.7. Survival curves
Supplementary Figures 1–3 and Fig. 5 present survival curves for intravenous antibiotic treatment, rounded abscess morphology with mass effect, and abscess long-axis thresholds of 30 mm and 50 mm. These thresholds were used for Kaplan-Meier visualization only, whereas the Cox models analyzed long axis as a continuous variable. None of these variables demonstrated statistically significant differences in surgery-free survival at one year on Kaplan-Meier analysis. Percutaneous drainage was associated with a lower probability of remaining surgery-free in unadjusted Kaplan-Meier analysis (log-rank p = 0.03; Supplementary Figure 4). However, drainage was more commonly performed for large, well-defined collections with mass effect. After adjustment for collection size and morphology (Supplementary Figure 5), the association was attenuated and did not reach statistical significance, supporting a likely confounding-by-indication effect.
Fig. 5.
Kaplan–Meier survival curve for time to surgery based on abscess long axis diameter (threshold: 50 mm). Kaplan-Meier curve showing surgery-free survival according to abscess long-axis diameter using a 50-mm threshold. The p-value corresponds to the log-rank test.
4. Discussion
In our study, the risk factors for abdominal surgery at one year in Crohn's disease patients with intra-abdominal abscesses included abscess size and patient-specific characteristics evident on initial diagnostic imaging. Larger abscess size was the most consistent predictor of surgical risk. Patient age was also associated with a slightly elevated risk of surgery, while age at Crohn's disease diagnosis was inversely associated with surgical risk, suggesting a potential link between earlier disease onset and a higher likelihood of surgery. Rounded abscess morphology showed an association with reduced surgical risk compared to infiltrative abscesses, while intravenous antibiotic treatment was associated with a slight reduction in surgical risk.
Several studies have explored the clinical and radiological parameters linked to the response to medical treatment of intra-abdominal abscesses and the risk of surgery. For instance, Perl et al. [18] reported that over one-third of patients managed non-operatively for intra-abdominal abscesses avoided surgery for recurrent abscesses, disease recurrence, or medically refractory Crohn’s disease over a follow-up period of at least two years. Their study identified risk factors for failure of medical treatment, including larger abscess size (>6 cm), increased bowel wall thickness and the presence of small bowel dilatation detected by CT or MRI.
Casas Deza et al. [13] emphasize the importance of taking abscess size into account in the management of intra-abdominal abscesses. They propose antibiotic treatment for intra-abdominal abscesses smaller than 30 mm, and recommend surgical drainage for abscesses larger than 50 mm, with percutaneous drainage as an excellent alternative due to its favorable efficacy-safety ratio. Their data also suggest that patients treated with anti-TNF without bowel resection have a recurrence-free survival similar to that of patients who underwent surgery, aligning with our findings. Finally, a strong point of their study is its multicentric nature.
Yaari et al. [5] found the presence of stenosis as the only radiological factor independently associated with an increased risk of surgery at 1 year in patients with penetrating Crohn's disease. This finding contrasts with our results and those of Casas Deza et al. [13], in which stenosis was not predictive of surgical outcomes. A likely explanation is the difference in patient selection: Yaari et al. included exclusively patients with intra-abdominal fistulas, while our analysis specifically examined patients with intra-abdominal abscesses, reflecting distinct clinical populations.
Consistent with previous literature, particularly Perl et al. and Casas Deza et al., our multivariate analyses using Cox proportional hazards modeling confirmed abscess size as a key predictor of surgical risk. Unlike the simple resolution-based criteria used by Casas Deza et al., our methodology allowed simultaneous evaluation of multiple radiological, clinical, and biological variables, providing deeper insights into independent predictors of delayed surgery. Our variance and survival analyses further corroborated the relationship between larger abscess size and increased surgical risk, supporting the clinical relevance of this parameter.
Unlike the study by the GETECCU group, we did not find an association with maintenance therapies, particularly immunosuppressive or corticosteroid treatments. For corticosteroids, the lack of association could potentially be explained by the limited power of our study. However, the same reasoning does not apply to immunosuppressive treatments [13]. The age at Crohn's disease diagnosis is an interesting factor. Late-onset Crohn's appears to be associated with a lower risk of requiring surgery, which contrasts with the positive association of age itself observed in our multivariate model. These findings align with trends related to disease duration, which suggest a higher likelihood of surgery over time, as reported in studies by Perl and the GETECCU group. [13], [18], [19]. Notably, variables such as age at diagnosis, infiltrative collections, and IV antibiotic use were not significant in univariate analysis and in Kaplan–Meier survival analyses but reached significance in the multivariate model, likely reflecting confounding or intervariable correlations. These findings highlight the strength of multivariate modeling in uncovering independent predictors within complex clinical data.
Percutaneous drainage deserves specific caution in interpretation. In our cohort, drainage was more common among patients who ultimately underwent surgery, but this observation is highly susceptible to confounding by indication, because drainage is preferentially performed in larger, well-defined, technically accessible collections and in clinically more severe presentations. These findings are comparable to those of Gervais et al., who reported failures requiring surgical intervention [20], [21]. Therefore, percutaneous drainage was not included as a baseline covariate in the main multivariate model, as it likely represents a consequence of initial abscess characteristics and clinical decision-making rather than an independent prognostic factor. Our results should not be interpreted as evidence that drainage itself increases surgical risk.
One key limitation of our study is its retrospective and single-center design. Patient management may have been influenced by variations in clinical practices, patient preferences, and the absence of a uniform written therapeutic protocol throughout the study period. Although patients were managed in a tertiary IBD referral center and were generally discussed by gastroenterologists, surgeons, radiologists, and interventional radiologists when drainage or surgery was considered, medical management, drainage indications, and timing of surgery were not standardized. This may have introduced treatment-selection bias, particularly for percutaneous drainage. In addition, patients were initially identified using ICD codes. Although each included case was individually reviewed to confirm eligibility, this strategy may have missed patients with absent or inaccurate coding and may therefore have introduced selection bias.
Other limitations relate to the definition of the study outcome and the duration of follow-up. Delayed surgery was defined as surgery occurring between day 15 and one year after abscess diagnosis. The lower threshold was selected pragmatically to exclude patients requiring immediate or emergency surgery and to focus on a clinically relevant period during which predictive factors may help guide management decisions. The one-year follow-up period also remains debatable. Surgery beyond one year may be influenced by recurrent abscess, fibrostenotic progression, medically refractory disease, or new penetrating complications rather than the index abscess itself. However, most events occurred within the first 150 days, as shown by the survival curves. Although follow-up imaging around day 40 was available for most non-operated patients, it was not systematically available for all patients. Therefore, one-year surgery-free status was based on clinical follow-up and the absence of surgery, rather than on systematic radiological confirmation of complete abscess resolution. In addition, detailed post-abscess or postoperative therapeutic sequences, including the exact timing of biologic initiation or reintroduction, could not be consistently analyzed.
Another limitation is the moderate interobserver agreement observed for certain imaging characteristics, particularly enhancing abscess wall assessment and fistula type. Discontinuous or partially visible walls may be interpreted differently among readers. Similarly, distinguishing between a sinus tract, a simple and a complex fistula can be challenging to determine reliably, particularly in cases of atypical anatomy, multiple tracts, or poorly defined extensions. However, these levels of agreement align with those reported in the literature [22], [23], suggesting that the findings remain broadly reliable. It is important to note that our kappa coefficients were calculated without distinguishing between imaging modalities (CT and MRI), unlike prior studies. For instance, Jensen and Schmidt reported better reproducibility with CT compared to MRI for detecting Crohn's disease and digestive parietal changes, with κ values around 0.65 for CT and 0.48 for MRI [23], [24]. Conversely, Schleder found good reproducibility with MRI for identifying inflammatory abnormalities of the ileum, with Kendall's coefficients of 0.724 for the ileum and 0.823 for the terminal ileum [25].
In conclusion, in patients with Crohn’s disease presenting with an intra-abdominal abscess, larger abscess long-axis diameter and older age were associated with a higher risk of delayed surgery between day 15 and one year. Conversely, later age at Crohn’s disease diagnosis, well-circumscribed morphology with mass effect, and intravenous antibiotic therapy were associated with a reduced risk of delayed surgery. These findings should be interpreted as exploratory and may help generate hypotheses for future prospective studies aiming to tailor individualized management strategies.
Ethics
This retrospective study was approved by the Comité d′Éthique pour la Recherche en Imagerie Médicale (CERIM) of the Collège des Enseignants en Radiologie de France (CERF), under IRB number CRM−2404–404.
Statements and Declarations
The authors declare that they have no competing interests.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
CRediT authorship contribution statement
Lionel Arrive: Writing – review & editing, Supervision, Project administration, Methodology, Investigation. Pauline Iorio: Visualization, Validation, Resources, Investigation, Formal analysis. Philippe Seksik: Writing – review & editing, Resources, Project administration, Funding acquisition. Harry Sokol: Writing – review & editing, Validation, Resources. Julien Kirchgesner: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Formal analysis, Conceptualization. Caroline Su: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Quentin Vanderbecq: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Mathilde Wagner: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Formal analysis.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ejro.2026.100802.
Appendix A. Supplementary material
Supplementary material
References
- 1.Thia K.T., Sandborn W.J., Harmsen W.S., Zinsmeister A.R., Loftus E.V. Risk factors associated with progression to intestinal complications of Crohn’s disease in a population-based cohort. Gastroenterology. 2010;139:1147–1155. doi: 10.1053/j.gastro.2010.06.070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Feagins L.A., Holubar S.D., Kane S.V., Spechler S.J. Current Strategies in the Management of Intra-abdominal Abscesses in Crohn’s Disease. Clin. Gastroenterol. Hepatol. 2011;9:842–850. doi: 10.1016/j.cgh.2011.04.023. [DOI] [PubMed] [Google Scholar]
- 3.Adamina M., Bonovas S., Raine T., Spinelli A., Warusavitarne J., Armuzzi A., Bachmann O., Bager P., Biancone L., Bokemeyer B., Bossuyt P., Burisch J., Collins P., Doherty G., El-Hussuna A., Ellul P., Fiorino G., Frei-Lanter C., Furfaro F., Gingert C., Gionchetti P., Gisbert J.P., Gomollon F., González Lorenzo M., Gordon H., Hlavaty T., Juillerat P., Katsanos K., Kopylov U., Krustins E., Kucharzik T., Lytras T., Maaser C., Magro F., Marshall J.K., Myrelid P., Pellino G., Rosa I., Sabino J., Savarino E., Stassen L., Torres J., Uzzan M., Vavricka S., Verstockt B., Zmora O., European Crohn’s and Colitis Organisation [ECCO] ECCO Guidelines on Therapeutics in Crohn’s Disease: Surgical Treatment. J. Crohns Colitis. 2020;14:155–168. doi: 10.1093/ecco-jcc/jjz187. [DOI] [PubMed] [Google Scholar]
- 4.Bossuyt P., Debeuckelaere C., Ferrante M., Vanbeckevoort D., Billiet T., Wolthuis A., van Assche G., D’Hoore A., Vermeire S. The operative risk and natural history after the diagnosis of ileal penetrating Crohn’s disease. Eur. J. Gastroenterol. Hepatol. 2018;30:539–545. doi: 10.1097/MEG.0000000000001091. [DOI] [PubMed] [Google Scholar]
- 5.Yaari S., Benson A., Aviran E., Lev Cohain N., Oren R., Sosna J., Israeli E. Factors associated with surgery in patients with intra-abdominal fistulizing Crohn’s disease. World J. Gastroenterol. 2016;22:10380. doi: 10.3748/wjg.v22.i47.10380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Malgras B., Soyer P., Boudiaf M., Pocard M., Lavergne-Slove A., Marteau P., Valleur P., Pautrat K. Accuracy of imaging for predicting operative approach in Crohn’s disease. Br. J. Surg. 2012;99:1011–1020. doi: 10.1002/bjs.8761. [DOI] [PubMed] [Google Scholar]
- 7.Fallis S.A., Murphy P., Sinha R., Hawker P., Gladman L., Busby K., Sanders S. Magnetic resonance enterography in C rohn’s disease: a comparison with the findings at surgery. Colorectal Dis. 2013;15:1273–1280. doi: 10.1111/codi.12361. [DOI] [PubMed] [Google Scholar]
- 8.Bruining D.H., Zimmermann E.M., Loftus E.V., Sandborn W.J., Sauer C.G., Strong S.A., Al-Hawary M., Anupindi S., Baker M.E., Bruining D., Darge K., Dillman J., Einstein D., Fidler J.L., Fletcher J.G., Gee M.S., Grand D., Guglielmo F., Hara A.K., Jaffe T.A., Maglinte D., Megibow A., Park S.H., Platt J., Podberesky D., Rimola J., Sahani D.V., Soto J., Taylor S.A. Consensus Recommendations for Evaluation, Interpretation, and Utilization of Computed Tomography and Magnetic Resonance Enterography in Patients With Small Bowel Crohn’s Disease. Gastroenterology. 2018;154:1172–1194. doi: 10.1053/j.gastro.2017.11.274. [DOI] [PubMed] [Google Scholar]
- 9.Peyser D.K., Carmichael H., Dean A., Baratta V.M., D’Andrea A.P., Kang G., Bhasin D., Greenstein A.J., Khaitov S.K., Steinhagen R.M., Sylla P. Early versus delayed ileocolic resection for complicated Crohn’s disease: is “cooling off” necessary? Surg. Endosc. 2022;36:4290–4298. doi: 10.1007/s00464-021-08773-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Nguyen D.L., Sandborn W.J., Loftus E.V., Larson D.W., Fletcher J.G., Becker B., Mandrekar J., Harmsen W.S., Bruining D.H. Similar Outcomes of Surgical and Medical Treatment of Intra-abdominal Abscesses in Patients With Crohn’s Disease. Clin. Gastroenterol. Hepatol. 2012;10:400–404. doi: 10.1016/j.cgh.2011.11.023. [DOI] [PubMed] [Google Scholar]
- 11.Gionchetti P., Dignass A., Danese S., Magro Dias F.J., Rogler G., Lakatos P.L., Adamina M., Ardizzone S., Buskens C.J., Sebastian S., Laureti S., Sampietro G.M., Vucelic B., van der Woude C.J., Barreiro-de Acosta M., Maaser C., Portela F., Vavricka S.R., Gomollón F. on behalf of ECCO, 3rd European Evidence-based Consensus on the Diagnosis and Management of Crohn’s Disease 2016: Part 2: Surgical Management and Special Situations. J. Crohns Colitis. 2017;11:135–149. doi: 10.1093/ecco-jcc/jjw169. [DOI] [PubMed] [Google Scholar]
- 12.Bouhnik Y., Pineton de Chambrun G., Lambert J., Nachury M., Seksik P., Altwegg R., Vuitton L., Stefanescu C., Nancey S., Aubourg A., Serrero M., Filippi J., Desseaux K., Viennot S., Abitbol V., Boualit M., Bourreille A., Giletta C., Buisson A., Roblin X., Dib N., Malamut G., Amiot A., Fumery M., Louis E., Elgharabawy Y., Peyrin-Biroulet L., Moreau J., Treton X., Mailhat C., Pollet J., Martins C., Savoye G., Tavernier N., Mary J.-Y., Zappa M., Painchart C. Adalimumab in Biologic-naïve Patients With Crohn’s Disease After Resolution of an Intra-abdominal Abscess: A Prospective Study From the GETAID. Clin. Gastroenterol. Hepatol. 2023 doi: 10.1016/j.cgh.2023.01.013. [DOI] [PubMed] [Google Scholar]
- 13.Casas Deza D., Polo Cuadro C., De Francisco R., Vela González M., Bermejo F., Blanco I., De La Serna Á., Bujanda L., Bernal L., Rueda García J.L., Gargallo-Puyuelo C.J., Fuentes-Valenzuela E., Castro B., Guardiola J., Ladrón G., Suria C., Sáez Fuster J., Gisbert J.P., Sicilia B., Gomez R., Muñoz Vilafranca C., Barreiro-De Acosta M., Peña E., Castillo Pradillo M., Cerrillo E., Calvet X., Manceñido N., Monfort I Miquel D., Marín S., Roig C., Marce A., Ramírez De Piscina P., Betoré E., Martin-Cardona A., Teller M., Alonso Abreu I., Maroto N., Frago S., Gardeazabal D., Pérez-Martínez I., Febles González Á.D., Barrero S., Taxonera C., García De La Filia I., Ezkurra-Altuna A., Madero L., Martín-Arranz M.D., Gomollón F., Domènech E., García-López S. Initial Management of Intra-abdominal Abscesses and Preventive Strategies for Abscess Recurrence in Penetrating Crohn’s Disease: A National, Multicentre Study Based on ENEIDA Registry. J. Crohns Colitis. 2024;18:578–588. doi: 10.1093/ecco-jcc/jjad184. [DOI] [PubMed] [Google Scholar]
- 14.Grellier N., Kirchgesner J., Uzzan M., Mclellan P., Stefanescu C., Lefèvre J.H., Treton X., Panis Y., Sokol H., Beaugerie L., ERIC Study group, Seksik P. Early ileal resection in Crohn’s disease is not associated with severe long-term outcomes: The ERIC study. Aliment. Pharmacol. Ther. 2024;60:1388–1397. doi: 10.1111/apt.18247. [DOI] [PubMed] [Google Scholar]
- 15.Le Baut G., Kirchgesner J., Amiot A., Lefevre J.H., Chafai N., Landman C., Nion I., Bourrier A., Delattre C., Martineau C., Sokol H., Seksik P., Nguyen Y., Marion Y., Lebreton G., Carbonnel F., Viennot S., Beaugerie L. Saint Antoine IBD Network, A Scoring System to Determine Patients’ Risk of Colectomy Within 1 Year After Hospital Admission for Acute Severe Ulcerative Colitis. Clin. Gastroenterol. Hepatol. Off. Clin. Pract. J. Am. Gastroenterol. Assoc. 2021;19:1602–1610.e1. doi: 10.1016/j.cgh.2019.12.036. [DOI] [PubMed] [Google Scholar]
- 16.Silverberg M.S., Satsangi J., Ahmad T., Arnott I.D., Bernstein C.N., Brant S.R., Caprilli R., Colombel J.-F., Gasche C., Geboes K., Jewell D.P., Karban A., Loftus E.V., Jr, Peña A.S., Riddell R.H., Sachar D.B., Schreiber S., Steinhart A.H., Targan S.R., Vermeire S., Warren B.F. Toward an Integrated Clinical, Molecular and Serological Classification of Inflammatory Bowel Disease: Report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can. J. Gastroenterol. Hepatol. 2005;19 doi: 10.1155/2005/269076. [DOI] [PubMed] [Google Scholar]
- 17.Heinze G., Wallisch C., Dunkler D. Variable selection – A review and recommendations for the practicing statistician. Biom. J. 2018;60:431–449. doi: 10.1002/bimj.201700067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Perl D., Waljee A.K., Bishu S., Higgins P.D.R., Wasnik A.P., Stidham R.W. Imaging Features Associated With Failure of Nonoperative Management of Intraabdominal Abscesses in Crohn Disease. Inflamm. Bowel Dis. 2019;25:1939–1944. doi: 10.1093/ibd/izz069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Waked B., Holvoet T., Geldof J., Baert F., Pattyn P., Lobatón T., Hindryckx P. Conservative management of spontaneous intra-abdominal abscess in C rohn’s disease: Outcome and prognostic factors. J. Dig. Dis. 2021;22:263–270. doi: 10.1111/1751-2980.12984. [DOI] [PubMed] [Google Scholar]
- 20.Pugmire B.S., Gee M.S., Kaplan J.L., Hahn P.F., Doody D.P., Winter H.S., Gervais D.A. Role of percutaneous abscess drainage in the management of young patients with Crohn disease. Pediatr. Radiol. 2016;46:653–659. doi: 10.1007/s00247-015-3533-3. [DOI] [PubMed] [Google Scholar]
- 21.Gervais D.A., Hahn P.F., O’Neill M.J., Mueller P.R. Percutaneous Abscess Drainage in Crohn Disease: Technical Success and Short- and Long-term Outcomes during 14 Years. Radiology. 2002;222:645–651. doi: 10.1148/radiol.2223010554. [DOI] [PubMed] [Google Scholar]
- 22.Ziech M.L.W., Bipat S., Roelofs J.J.T.H., Nio C.Y., Mearadji B., Van Doorn S., Spijkerboer A.M., Stoker J. Retrospective comparison of magnetic resonance imaging features and histopathology in Crohn’s disease patients. Eur. J. Radiol. 2011;80:e299–e305. doi: 10.1016/j.ejrad.2010.12.075. [DOI] [PubMed] [Google Scholar]
- 23.Schmidt S., Lepori D., Meuwly J.-Y., Duvoisin B., Meuli R., Michetti P., Felley C., Schnyder P., van Melle G., Denys A. Prospective comparison of MR enteroclysis with multidetector spiral-CT enteroclysis: interobserver agreement and sensitivity by means of “sign-by-sign” correlation. Eur. Radiol. 2003;13:1303–1311. doi: 10.1007/s00330-002-1710-x. [DOI] [PubMed] [Google Scholar]
- 24.Jensen M.D., Ormstrup T., Vagn-Hansen C., Østergaard L., Rafaelsen S.R. Interobserver and Intermodality Agreement for Detection of Small Bowel Crohn’s Disease with MR Enterography and CT Enterography. Inflamm. Bowel Dis. 2011;17:1081–1088. doi: 10.1002/ibd.21534. [DOI] [PubMed] [Google Scholar]
- 25.Schleder S., Pawlik M., Wiggermann P., Ott C., Fichtner-Feigl S., Müller-Wille R., Stroszczynski C., Schreyer A.G. Interobserver Agreement in MR Enterography for Diagnostic Assessment in Patients with Crohn’s Disease, RöFo - Fortschritte Auf Dem Geb. Röntgenstrahlen Bildgeb. Verfahr. 2013;185:992–997. doi: 10.1055/s-0033-1335445. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material