Percutaneous abdomino-pelvic abscess drainage in complicated Crohn’s disease

Crohn’s disease (CD) is a chronic, relapsing in lammatory disorder of the gastrointestinal tract characterized by a transmural in lammation of the bowel wall leading to complications including perforation, istula and abdominopelvic abscess onset. This latter event is present in a percentage of CD patients ranging between 20% 30% [1-6]. In the past, patients with an abdominopelvic abscess were treated with surgical drainage followed by bowel resection. Currently, percutaneous abscess drainage (PAD) is considered the standard therapy for abdominopelvic infected luid collections due to its signi icant reduction of morbidity and mortality compared to surgical procedures [1,7-14].


Introduction
Crohn's disease (CD) is a chronic, relapsing in lammatory disorder of the gastrointestinal tract characterized by a transmural in lammation of the bowel wall leading to complications including perforation, istula and abdominopelvic abscess onset. This latter event is present in a percentage of CD patients ranging between 20% -30% [1][2][3][4][5][6].
In the past, patients with an abdominopelvic abscess were treated with surgical drainage followed by bowel resection. Currently, percutaneous abscess drainage (PAD) is considered the standard therapy for abdominopelvic infected luid collections due to its signi icant reduction of morbidity and mortality compared to surgical procedures [1,[7][8][9][10][11][12][13][14]. and/or percutaneous drainage [15]. The safety and ef icacy of PAD as irst-line therapy has been demonstrated in several retrospective studies. In particular, studies reported high ef icacy of PAD in postoperative collections in CD while spontaneous abscesses are more resistant to PAD due to their frequent association to istulas and the underlying stenosis [4,7,9,16,17].
Nowadays, PAD has been employed with the goals of avoiding or better, delaying surgery until sepsis and clinical patients' conditions are resolved or improved. Technical approach and clinical success rates on patients' series were widely reported in the literature [18] however, studies on the timing to surgery and factors affecting the clinical success rate are limited [19,20].
We retrospectively evaluated both the ef icacy in terms of technical and clinical success rate, complications rate and inal outcome of PAD in patients with abscesses from CD considering separately postoperative and spontaneous abscesses.
We also analyzed the factors which can potentially in luence the ef icacy of the procedure.

Methods
The present study ful illed the Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data. The study conforms to the ethics guidelines of the Declaration of Helsinki and data collection and analyses had already previously been approved by Institutional Review Board of our hospital. Informed consent was obtained from patients who participated in clinical investigations. In the period between September 2006 and March 2019 at our Department we performed 114 PAD in patients with abscesses from CD. Six patients were excluded because lost at follow-up; in addition, we excluded 17 patients submitted to simple needle aspiration of luid collection < 3 cm in diameter; therefore, our inal study population resulted in 91 patients including 45 postoperative abscesses [occurred within 30 days post-operative (p.o.)] and 46 spontaneous "disease-related" abscesses ( Figure 1).
The inclusion criteria were luid infected collections > 3 cm in diameter located in upper (right and/or left hypochondrium) or lower abdomen/pelvis (pelvis; right and/or left iliac fossa) associated to con irmed diagnosis of CD and signs or symptoms of sepsis (fever, leukocytosis, local pain). Postoperative collections were considered any abscess >3 cm occurring within 30 days from the abdominal surgery and documented at imaging.
All relevant clinical information (demographics, disease location, previous medical treatments, surgical history) were collected from our institutional database.
Short and long-term follow-up information (extended to at least 3 years in all patients) to verify the clinical outcome after PAD were collected.
For our purpose we de ined the "Overall success" (OS) of the PAD as "Technical success" (TS) (complete resolution of the abscess documented by imaging) plus "Clinical success" (CS) (no need for surgery within the irst 30 days post-drainage due to defervescence, decreased white blood cell count and improvement of the clinical condition of the patient).
Patients who needed surgery within 30 days after the procedure but had no residual abscess at the time of surgery were considered as TS but clinical failure (CF).
Technical failure (TF) was de ined as the presence of a residual abscess. Overall Failure (OF) was de ined as CF with or without TF.
Complications of PAD were classi ied as major or minor according to Interventional Radiology Criteria [21].
Any coagulopathy (INR > 1.5, platelet count < 50.000/ mm3) was corrected and all the procedures were performed by an experienced interventional radiologist with at least 10 years of experience.
After the placement of the catheter, systemic antibiotic therapy was given until the signs and symptoms of the infection abated (reduction of fever and leukocytosis) and catheters were kept patent by irrigation with 10-20 ml saline solution two to four times a day, depending on the viscosity of the luid drained. Patients were evaluated for catheter-related complications and/or istulous communication with ultrasonographic (US; Esaote), computed tomography (CT; GE Healtcare Lighspeed VCT) or trans-drainage istulography CT weekly based on clinical and radiological feature.
All catheters have been removed when an amount of daily drainage less than 10 mL/24 h was observed, and when providing it with a CT scan that con irmed a complete abscess resolution.
After the removal of the catheter, patients were followed with MR enterography (MR; GE Signa 1.5 T) yearly to evaluate both the evolution of the disease and the presence of istulas and the relapse of the abscess.

Statistical analysis
All the analyses were performed using SPSS vs. 20 software. Categorical variables were expressed as frequencies and percentages, while the continuous ones as means and ranges. The signi icance of the differences between groups for continuous variables was assessed with the Student t test, whereas the chi square test of Pearson was applied for percentages. Signi icance was set at p < 0.05.

Results
Main demographic and clinical characteristics of the study population are shown in table 1. The mean age of the 91 patients was 39 years (range 15-65 years) and 53 male and 38 female.
The mean diameter of the abscesses was 6.5 cm, ranging from 3 to 30 cm.  PAD was performed under CT guidance in 86 (94%) patients and US guidance in the remaining 5 (6%) cases. Drainage placement was performed without any major complications (septic shock, bacteremia, profuse bleeding requiring transfusion or bowel perforation) in all 91 patients. Only 1 patient experienced a minor complication (haematoma of the abdominal wall) resolved conservatively.
In the postoperative group we had OS 41 (91%), OF 4 (9%) and no TF; the global rate of technical success was 100%.
In the spontaneous group we had OS 15 (33%), OF 31 (67%) and TF 2 (6.4%); the global rate of technical success was 44 (95,6%). Therefore, a total abscess drainage with resolution of the sepsis was achieved in the 89 (97,8%) of the whole study population.
The outcome of PAD is summarized in igure 3 and table 2.
We also analyzed which factors were potentially correlated with the outcome of the procedure ( Table 2). Postoperative abscesses have been successfully resolved more frequently than spontaneous ones (p < 0.0001). We found no statistically signi icant differences among sex, age, abscess size and disease duration.
Thirty-two (35.2%) of OS and 12 (13.2%) failures were achieved in the upper abdomen while 24 (26.4%) successes ad 23 (25.2%) failures in the lower abdomen/pelvis; these results to be statistically signi icant (p < 0.05) for a best outcome of upper location.
The characteristics of the abscess (multiloculated . uniloculated) were found to be statistically signi icant related to a predictor of success after PAD.
The mean length of hospitalization was 26 days (range 1-116 days) and the duration of drainage was of 15 days (range 7-120 days).
Interestingly, most of all the postoperative collections did not need early surgery; on the contrary, patients with a spontaneous abscess had an extremely variable clinical picture, regardless of the effectiveness of the drainage. Bowel resection was necessary within 1 month mainly due to obstructive symptoms, the coexistence of multiple istulas, the presence of associated mesenterial phlegmons and poor quality of life in the patients. The substantial difference between the two subgroups was also evidenced by the discrepancy in the average length of hospitalization after drainage for the postoperative and spontaneous group (11 vs. 21 days respectively).
We also analyzed the in luence of the pharmacological history of patients on the effectiveness of PAD.
Due to the dif iculties in order to stratify patients on the basis of the therapies, we focused our attention on the use of biological drugs within 3 months from the onset of the abscess. Eleven out of 91 (12%) patients had biological drugs in their pharmacological history, consequently the sample's meanness did not allow to draw meaningful conclusions.

Discussion
Percutaneous drainage is an effective and safe procedure for CD complicated by an abdominopelvic abscess. The more recent ECCO-ESCP guidelines state that intra-abdominal abscess should initially be treated with antibiotics and/or percutaneous drainage and the surgical approach should be limited to the emergent cases [15]. Due to the extreme variability of the characteristics of collections patients and variability of CD itself, that can in luence the success of the procedure, the curative rate of the drainage changes signi icantly between studies (38% -100%) [8,24,25].
However, the ef icacy of PAD is reported to be high for postoperative collections, while spontaneous abscesses are more resistant to PAD due to their frequent multiloculated features and the association with advanced postin lammatory damages of the bowel wall mainly stenosis and istulas [1,3,9,[26][27][28].
Nowadays, it is well established that PAD is useful also in a preoperative setting in order to improve patients' clinical conditions for further elective surgical procedure or better to avoid surgery.
However, its role in clinical management in severe complicated CD patients who need an early surgery, is not well established.
Data from different outcomes in postoperative and spontaneous abscesses which are different in the outcome due to the underlying clinical condition, are not available.
Our irst goal was to retrospectively review the ef icacy and safety of PAD both in postoperative and spontaneous abscess (this latter group included patients who needed an early surgery within 30 days).
Our main goal was to determine the rate of the OS of PAD de ined as avoidance of surgery due to clinical improvement of the patients (CS) and to complete the resolution of the abscess (TS).
However, we added an additional criterion to better de ine the usefulness of PAD that is the clinical failure that means patients who needed surgery within 30 days from PAD due to the underlying severe disease but had a complete resolution of the abscess (TS).
This criterion was included because it is well known that PAD is used as a bridge to elective surgical bowel resection for severe persistent diseases complicated by strictures and istulas.
In order to better discriminate our results, we considered data from postoperative and spontaneous abscesses separately.
Using these criteria, we found a global rate of technical success of 100% in the postoperative setting and 95.6% in the spontaneous group.
Our data is similar to those reported in the literature: in fact, it is well known that predictors of poor outcomes after PAD as a curative treatment are usually considered abscess characteristics (aetiology in terms of spontaneous or postoperative, unilocular or multilocular collections, location, size, number and presence of istulas) [9,16,24,27,[29][30][31][32].
We found a statistically signi icant correlation with a better outcome after PAD in patients with postoperative collections; one of the reasons may be the signi icant lower rate of multiloculated collections in this group that is 4% vs. 39% of the spontaneous group. This latter characteristic of the abscess was related to a better outcome after PAD with a reported failure in 38% of the cases vs. 1% of the uniloculated ones (p < 0.0001). Even though we found quite a similar number of documented istulas in both Groups, the differences in istula tract was evident particularly in the postoperative patients, istulas were mainly due to an anastomotic dehiscence and the collection originated from a peritoneal contamination occurred at the time of surgery. On the other side in the spontaneous group istulas are mainly related to severe intestinal stenosis which causes istulas, deep intestinal wall ulcerations and poor clinical conditions; this condition makes it hard to achieve a high clinical success in this population. Akinci, et al. [20]. reported the presence of istulas as the only factor affecting the clinical success rate after PAD in a group of 185 pelvic abscesses. Other authors reported that predictors of unsuccessful outcomes after PAD are abscesses caused by internal wide istulas which is an indication for surgical repair in a range of 8%-18% of the cases [30,33,[34][35][36].
Differently from other experiences in our study population, we did not ind a strong difference in the OS on the basis of the collection location reporting a trend in the signi icance between success and failure in patients with a collection in the upper abdomen. A possible explanation could be the most complicated access routes through the abdominal wall, especially in postoperative patients; moreover, pelvic collections are often winding and supported by enterovesical, entero-vaginal and entero-sigmoid istulas (data non reported).
No statistically signi icant correlation was found between the outcome of the procedure and the assessed demographic variables such as gender, age of the patient at the time of hospitalization and previous years of illness.
Surprisingly, the size of the collection does not seem to in luence signi icantly the success of the procedure either: both abscesses between 3-6 cm and those beyond 7 cm have shown similar outcomes.
As reported from other authors [1,27] we observed collections up to 13 cm resolved with a PAD. In our experience we did not ind any major complications including septicemia with associated disseminated intravascular coagulation or hypotension, bowel perforation or death. Data from literature reports a complication rate both for spontaneous and postoperative abscesses in 8%-10% of the cases, mortality at 30 days ranging from 1% to 6% and puncturerelated mortality in around 0.7% [33,34,37] This variability may be due to varying levels of expertise in interventional radiology and the management of Crohn's patients in non-specialized centers.
Our better results may be referred to a preferential CT guided access technique due to the high number of deep collections.
The limitation of our paper is the retrospective evaluation, the limited number of patients and the absence of a strati ication for side, severity and location of the disease.
In fact we know that an abscess that is successfully drained, but results in a istula could not be considered a complete success of the drainage, but rather by delaying further surgery and allowing for patient optimization, therefore it is a success under a different light so we analyzed differently the technical and the clinical success.
In conclusion our data showed safety and effectiveness of PAD both in postoperative and spontaneous abscesses even in patients needing early surgery due to the severity of obstructive symptoms or active anastomotic istulas.