Diagnostic yield of cystography after sigmoid resection for colovesical fistula due to complicated diverticulitis.

Between 2010 and 2023, 204 patients underwent elective sigmoid resection for diverticulitis. In total 83 (40.7%) patients were male. The mean age was 62 years (± 10.6) for the whole cohort. In 55 (27.0%) of these patients a sigmoid resection was performed for a diverticular colovesical fistula (Fig.  1 ). In the fistula group, 37 (67.3%) patients were male ( p  ≤ 0.001). Mean age in the fistula group was 64 years (± 10.2) ( p  = 0.183). Preoperative symptoms experienced by patients with fistula were pneumaturia (83.6%), recurrent urinary tract infections (65.5%), and fecaluria (74.5%) (Table  1 ). Fig. 1 Patient inclusion and exclusion flowchart. Table 1 Baseline characteristics Elective sigmoid resections for diverticulitis n  = 204 Diverticulitis without fistula n  = 149 Diverticulitis with fistula n  = 55 P  value Age (years), mean ± SD 62 ± 10.6 61 ± 10.7 64 ± 10.2 0.183 Sex, male, n (%) 83 (40.7) 46 (30.9) 37 (67.3) < 0.001* BMI, kg/m 2 , mean ± SD 27 ± 4.7 27 ± 4.9 26 ± 4.2 0.318 ASA score, n (%)  I/II 165 (80.9) 123 (82.6) 42 (76.4) 0.319  III 39 (19.3) 26 (17.7) 13 (23.6) Smoking, n (%) 61 (29.9) 48 (32.2) 13 (23.6) 0.235 Diabetes mellitus, n (%) 19 (9.3) 13 (8.7) 6 (10.9) 0.634 Cardiovascular disease, n (%) 116 (56.9) 84 (56.4) 32 (58.2) 0.817 Previous abdominal surgery, n (%) 111 (54.4) 93 (62.4) 18 (32.7) < 0.001* Symptoms, n (%)  Pneumaturia 46 (83.6)  Recurrent urinary tract infections 36 (65.5)  Fecaluria 41 (74.5) Diagnostic imaging suspect for colovesical fistula, n (%)  CT scan 45/53 (84.9)  Cystoscopy 29/33 (87.8) BMI body mass index, ASA American Association of Anesthesiologists, CT computed tomography *Statistically significant Patient inclusion and exclusion flowchart. Baseline characteristics BMI body mass index, ASA American Association of Anesthesiologists, CT computed tomography *Statistically significant CT scan suggested the presence of a colovesical fistula in 45 out of 53 (84.9%) performed CT scans. Two patients did not undergo a diagnostic CT scan upon admission with characteristic symptoms. Recent CT scans in these patients had already shown diverticulitis, and cystoscopy confirmed fistula presence. Complementary to a CT scan, cystoscopy was performed if the patient first presented at the department of urology. A colovesical fistula was suspected in 29 out of 34 (85.5%) patients during cystoscopy (Table  1 ). Use of these diagnostic modalities allowed all fistulas to be diagnosed prior to surgery. In addition, patients with complicated diverticulitis underwent colonoscopy to rule out colorectal malignancy. However, colonoscopy was not used for diagnosis of a colovesical fistula. In the non-fistula group, 112 (81.8%) surgeries were performed using a minimally invasive approach (laparoscopic or robot assisted) compared to 40 (72.7%) surgeries in the fistula group ( p  = 0.833). The primary anastomosis rates were 93.3% and 88.9% for the non-fistula and fistula group, respectively ( p  = 0.304). Pelvic drains were placed based upon preference of the attending surgeon in 84 (56.4%) patients and 35 (63.6%) patients for the non-fistula and fistula group respectively ( p  = 0.351) (Table  2 ). In the fistula group, no drain-creatinine to serum-creatinine ratios were measured. Table 2 Surgical characteristics and outcomes Diverticulitis without fistula n  = 149 Diverticulitis with fistula n  = 55 P  value Approach, n (%)  Open 37 (24.8) 15 (27.3) 0.833  Laparoscopic 106 (71.1) 37 (67.3)  Robot assisted 6 (4.0) 3 (5.5)  Conversion 27 (18.1) 5 (9.1) 0.116 Duration of surgery (mins), mean ± SD 166 ± 61.6 173 ± 50.5 0.193 Procedure, n (%)  Primary anastomosis 139 (93.3) 48 (88.9) 0.304  Urologist participation in surgery, n (%) – 6 (10.9) Intraoperative bladder defect visualization, n (%) – 19 (34.5) Intraoperative vesical leakage testing, n (%) 2 (1.3) 19 (34.5) 0.778 Leakage during vesical leakage testing, n (%) – 2 (10.5) > 0.99 Closure of vesical defect, n (%)  Single layer defect closure – 18 (32.7) 0.351  Complex bladder repair 1 (0.7) 4 (7.3) Pelvic drain, n (%) 84 (56.4) 35 (63.6) Complications (Clavien–Dindo), n (%)  Grade I/II 32 (21.5) 16 (29.1) 0.489  Grade III/IV 8 (5.4) 2 (3.6) Anastomotic leakage, n (%) 2 (1.3) 1 (1.8) > 0.99 Re-intervention, n (%) 8 (5.4) 2 (3.6) > 0.99 Hospital length of stay, days, median [IQR] 4 [3–7] 5 [3–7] 0.040* *Statistically significant Surgical characteristics and outcomes *Statistically significant Vesical closure was performed in 22 out of 55 (40%) patients with a fistula. In 18 (32.7%) patients, the vesical closure was performed using a singular layer of sutures, utilizing monofilament, multifilament, or barbed sutures. In four (7.3%) patients a complex bladder repair was performed, defined as closure after partial bladder resection or closure requiring additional steps beyond the placement of a singular layer of sutures. The remaining 33 (60%) patients with fistula underwent no bladder repair, since the fistula was not identified during surgery by either visual inspection or methylene blue testing (Table  2 ). During surgery, no omental interpositions were applied between the colon and the bladder. In one patient without a colovesical fistula, partial bladder excision and complex bladder repair had been performed as a precaution, since there was no definitive conclusion about possible malignancy before surgery, and the process invaded into the bladder. Postoperative pathological analysis concluded the infiltrate was due to diverticulitis. Intraoperative methylene blue testing to assess vesical leakage was performed in 19 (34.5%) patients with colovesical fistula. No methylene blue leakage was observed in 17 out of 19 (89.5%) patients. In 2 out of 19 (10.5%) patients, methylene blue leakage was observed, after which simple suture repair was performed. Subsequent testing confirmed the absence of persisting methylene blue leakage. The complication rates between the non-fistula and fistula group were not significantly different, 26.9% versus 32.7% ( p  = 0.489). No complications related to IUC usage or bladder repair have occurred. The median length of hospital stay was significantly longer for the fistula group, 4 versus 5 days, ( p  = 0.040) (Table  2 ). Median IUC durations were 7 days (IQR 5–8) and 2 days (IQR 1–4) for the fistula and non-fistula group (p < 0.001). Cystography was performed in 37 (62.1%) of the patients with a colovesical fistula. Cystography was performed at a median of 7 days (IQR 5–8 days) after surgery. No extravesical contrast was observed on cystography in 34 (91.9%) of these patients. In three (8.1%) patients, extravesical contrast was observed on cystography. In two out of three patients extravesical contrast was detected 7 days after surgery after complex bladder repair which was performed by a urologist. Cystography was repeated after 7 and 14 days, respectively, revealing no contrast leakage. Consequently, the IUC was safely removed without complications. The third patient with abnormal cystography results 7 days after surgery showed a small fistula tract to an extravesical contrast pocket postoperatively. Intraoperative methylene blue testing (240 mL) revealed no contrast leakage; therefore, the fistula was not closed. Cystography was repeated after 14 days revealing no contrast leakage. Consequently, the IUC was safely removed without complications. Among the 17 patients without intraoperative methylene blue leakage, postoperative cystography demonstrated no extravesical contrast in 16 cases (94.1%). In the two patients who exhibited intraoperative methylene blue leakage and underwent subsequent simple suture repair, postoperative cystography showed no evidence of extravesical contrast. Overall, negative cystography results were observed in 94.7% of cases without intraoperative leakage or following simple suture repair. Moreover, in all 18 patients who did not undergo cystography prior to IUC removal, no symptoms developed necessitating cystography or IUC replacement. The median IUC duration did not differ significantly between patients who underwent cystography and those who did not, 7 versus 6 days ( p  = 0.104). During follow-up, no significant difference in long-term lower urinary tract symptoms were observed in these groups, 9 versus 2 cases ( p  = 0.203). Colovesical fistula recurrence was not detected during follow-up (Table  3 ). Table 3 Postoperative cystography and urological characteristics in the fistula group IUC removal after cystography n  = 37 IUC removal without cystography n  = 18 P  value Time to cystography, days, median [IQR] 7 [5–8] – Extravesical contrast observed on cystography, n (%) 3 (8.1) Time to IUC removal, days, median [IQR] 7 [5–9] 6 [5–7] 0.104 Urological complications, n (%)  LUTS during hospital stay 2 (5.4) 2 (11.1) 0.251  LUTS during follow-up 9 (24.3) 2 (11.1) 0.203 Hospital stay, days, median [IQR] 4 [3–7] 6 [5–7] 0.156 Colovesical fistula recurrence, n (%) – – Follow-up, days, median [IQR] 50 [26–110] 46 [30–132] 0.936 IUC indwelling urinary catheter, LUTS lower urinary tract symptoms Postoperative cystography and urological characteristics in the fistula group IUC indwelling urinary catheter, LUTS lower urinary tract symptoms

Postoperative cystography has limited diagnostic yield 7 days after surgery and may not be necessary in patients with diverticular colovesical fistula without or after simple bladder repair. However, cystography should be performed if complex bladder repair has been performed. These conclusions should be interpreted in the context of the study’s retrospective design and its limitations. Prospective studies are warranted to validate these findings and improve future clinical guidelines.

This study investigates the diagnostic yield of cystography after sigmoid resection for diverticular colovesical fistula and its influence on the postoperative course. In addition, a comparative analysis of diverticulitis with and without a colovesical fistula is conducted. Our study demonstrates that the diagnostic yield of routinely cystography performance 7 days after surgery, before IUC removal following sigmoid resection for a colovesical fistula is low. Furthermore, surgery for diverticular colovesical fistula is associated with longer hospital stay compared with sigmoid resection for diverticulitis without a fistula. In addition, surgery for diverticular colovesical fistula is not associated with an otherwise abnormal (post)operative course. At our center, 37 out of 55 (67.3%) patients with fistula underwent postoperative cystography before IUC removal. Although the local protocol advised to perform cystography before IUC removal, protocol violations were frequently observed. We hypothesize that this might be related to the relatively large number of surgeons performing these procedures at a low frequency in the beginning of the study period. Later on, this procedure was performed by three dedicated colorectal surgeons. From this group, in 34 out of 37 (91.9%) patients who underwent cystography, no vesical leakage was found on cystography. In three (8.1%) patients with fistula, extravesical contrast was observed on cystography. In two of these three patients, the extravesical contrast was observed in patients who underwent a complex bladder repair. In one of these two patients, intraoperative vesical leakage testing showed methylene blue leakage. In addition, in one patient cystography revealed a small remaining fistula tract; however, no contrast leakage was observed during intraoperative testing. Since only 240 mL of methylene blue was injected into the bladder, we believe that this volume may have been insufficient to reveal the vesical defect. These results correspond to findings in the literature in which high rates of negative cystography findings are presented, ranging from 80% to 100% [ 10 – 14 , 19 ]. Even so, not all patients undergoing sigmoid resection for colovesical fistula require cystography before IUC removal, as confirmed by our study [ 10 , 12 , 13 , 18 , 19 ]. The current literature lacks references regarding the diagnostic benefit (sensitivity and specificity) of cystography for this indication. Cystography usage differs greatly between hospitals and surgeons [ 5 , 9 – 12 , 14 , 16 , 18 – 20 ]. Several studies suggest cystography is only of benefit in patients with impaired vesical healing, e.g., if a large vesical defect is present, after complex bladder repair, after radiation in the surgical area, and if the overall condition of the patient is frail [ 10 , 12 – 14 , 18 , 19 ]. In contrast, other studies state that postoperative cystography should not be performed routinely [ 11 , 22 – 24 ]. Multiple studies removed IUC between 7 to 10 days after surgery without prior cystography performance [ 22 – 24 ]. In none of the studied patients were complications such as intra-abdominal urinary leakage or fistula recurrence observed after IUC removal [ 22 – 24 ]. Instead, the authors state that a prolonged IUC duration of 7 to 14 days is a sufficient postoperative policy [ 11 , 22 – 24 ]. Median IUC duration in our center was 7 days for all patients with colovesical fistula. In a recent study, Stapler et al. concluded that IUC removal at 1 to 2 days postoperative is safe for small vesical defects in patients without severe comorbidities or other reasons to leave the IUC for a longer period [ 12 ]. IUC removal on day 1 to 2 was not related to a rise in postoperative complications [ 12 ]. No data is available about the frequency of cystography performance in these patients. However, in all 55 performed cystographies for patients with less than 7 days of IUC duration, no leakage was observed on cystography. In addition, results from a single-arm prospective trial indicate that IUC removal on postoperative day 2 to 3 is feasible and safe after obtaining negative cystography, demonstrating 92% negative cystography results on postoperative day 2 [ 14 ]. Prolonged IUC duration (> 7 days) may be indicated on the basis of the condition of the patient or if impaired recovery after surgery is expected [ 12 ]. To further improve postoperative IUC duration and cystography performance, some studies state that postoperative cystography is not necessary when intraoperative vesical leakage testing reveals no leakage [ 10 , 12 , 14 ]. This strategy has also shown favorable results in evaluation of the integrity of vesicourethral anastomosis [ 25 ]. In our cohort, vesical leakage tests revealed no methylene blue leakage and subsequent extravesical contrast on cystography in 94.7% of cases, comparable to findings in the literature [ 12 , 13 ]. In one (5.3%) patient, however, extravesical contrast was detected on cystography after methylene blue leakage had not been observed during surgery. Moreover, intraoperative vesical leakage testing has been associated with a reduction in vesical repairs and a shorter IUC duration in cases with negative test results [ 12 ]. Next to intraoperative vesical leakage testing, Dolejs et al. suggest that patients with positive test results—particularly those at highest risk—should be considered for closed-suction drainage. In addition, they propose using postoperative drain creatinine-to-serum creatinine ratios to detect vesical leakage. In their study, this ratio was determined in 18 patients ratios and demonstrated a sensitivity of 75%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 93%. On the basis of those findings, they consider it a cost-effective and less invasive method to exclude or confirm vesical leakage [ 10 ]. This test has not been implemented at our center. Further research is needed to validate its utility, particularly in high-risk patients. On the basis of the outcomes of our study and a combination of the findings in the literature presented above, we offer the algorithm for peri- and postoperative colovesical fistula care presented in Fig.  2 , which can be used in daily clinical practice. To start, intraoperative vesical leakage testing using a solution containing methylene blue can play a role in identifying the leak and closing it. We suggest injecting at least 500 mL or until the bladder is at a substantial tension to ensure there is sufficient pressure on the bladder to detect even small defects. In addition, we suggest that the IUC may be removed on postoperative day 3 without performing cystography if no urinary leakage is observed during intraoperative bladder leakage testing [ 10 , 12 , 14 ]. If leakage is observed and the vesical defect can be closed with a simple suture repair without subsequent methylene blue leakage, the IUC may also be removed on postoperative day 3 without performing cystography. In cases of recurrent methylene blue leakage, we suggest reclosure of the defect followed by additional leakage testing to ensure adequate repair. Fig. 2 Algorithm for intra- and postoperative colovesical fistula care. . IUC indwelling urinary catheter. Algorithm for intra- and postoperative colovesical fistula care. . IUC indwelling urinary catheter. Extravesical contrast was observed on cystography in 50% of complex bladder repairs after 7 days, suggesting an IUC duration of 7 days may be needed in this group. However, as the complex group is very heterogeneous, patients may be present who are sufficiently treated with < 7 days of IUC usage. Therefore, we propose to perform cystography after 5–7 days in patients in whom complex bladder repair has been performed or delayed vesical recovery is expected, in accordance with recommendations by Stapler et al. [ 12 ]. In addition, this timeframe aligns with suggestions in urologic guidelines for (iatrogenic) vesical defects [ 26 ]. This algorithm offers potential for further optimization of Enhanced Recovery After Surgery (ERAS) protocols by decreasing the frequency of cystographies, as well as reducing the duration of IUC and hospitalization. A prospective study for validation of this algorithm is needed. Strengths of this study include, first, that it presents a consecutive cohort of patients reflecting daily care in a teaching hospital. Second, this study is, to our knowledge, the first study to provide a comparison between diverticulitis cases with and without a diverticular colovesical fistula. However, there are some limitations on this study. First of all, this is a retrospective uncontrolled study conducted at a single hospital which is subject to biases such as confounding, selection, and information bias. However, our series provides a sample size comparable to that of other studies, supporting the applicability of the findings. In addition, significant variability has been observed in (post)operative treatment. While this limits the generalizability of the results, it also underscores the potential benefit of the proposed treatment algorithm.

Diverticulosis and diverticulitis are common non-malignant diseases of the colon [ 1 ]. In the majority of patients, diverticulitis has an uncomplicated course. Only 12% experience complications, such as perforation, abscess formation, stenosis, or colovesical fistula [ 2 ]. Among patients with chronic and complicated diverticulitis, approximately 4% will develop a colovesical fistula [ 3 , 4 ]. A colovesical fistula involves a connection between the colon and the bladder. Patients with colovesical fistula usually present with recurrent urinary tract infections (UTI), pneumaturia, and/or fecaluria [ 5 ]. The current definitive treatment for diverticular colovesical fistula is a sigmoid resection. The morbidity rate following elective sigmoid resection for diverticulitis ranges from 7.3% to 28% [ 6 , 7 ]. If diverticulitis is complicated by a colovesical fistula, however, the morbidity rate ranges from 8.2% to 49.1% [ 5 ]. A meta-analysis comprising 127,169 elective sigmoid resections for diverticulitis revealed a 30-day mortality rate of approximately 0.5% [ 8 ]. In cases where diverticulitis is complicated by colovesical fistula, the mortality rate following sigmoid resection was approximately 1.5% in a series of 1061 patients [ 9 ]. Such outcomes may reflect the complexities associated with sigmoid resection for colovesical fistulas, emphasizing the need for structured guidelines in peri- and postoperative care. Despite this need, there is an absence of well-defined protocols when considering surgery for diverticular colovesical fistula. During surgery, the bladder defect may be closed with sutures. However, the fistula tract is frequently not identified and therefore not closed [ 5 , 9 – 11 ]. To identify the vesical defect, the bladder can be filled with a solution containing methylene blue or alternative contrast agents. Using methylene blue is thought to detect vesical leakage more easily. Performance of intraoperative leakage testing differs greatly between studies, ranging from 0% to 100% [ 10 , 12 – 16 ]. After surgery, it is common for patients to have an indwelling urinary catheter (IUC) to allow the bladder to heal under low intravesical pressure [ 5 , 10 , 17 ]. Before IUC removal, cystography can be performed to rule out persistent vesical leakage or a residual fistula tract. However, the use and timing of cystography as well as duration of IUC usage differ greatly between hospitals and surgeons [ 5 , 9 , 10 , 12 , 14 , 18 – 20 ]. The aim of this study was to evaluate the diagnostic yield of cystography after sigmoid resection for colovesical fistula and its clinical consequences. Moreover, differences in outcomes of sigmoid resection with and without a colovesical fistula were studied.

A retrospective single-center cohort study was conducted, including all patients who underwent elective sigmoid resection for complications related to diverticulitis (recurrent attacks, colovesical fistula, and stenosis) between 2010 and 2023. The study was performed at Isala Hospital in the Netherlands, a non-academic teaching hospital, in which over 400 colorectal procedures are performed annually. Included patients underwent elective surgery for diverticulitis and were at least 18 years of age at time of surgery. Patients with colovesical fistula due to malignancy, endometriosis, or prior bladder or rectosigmoid surgery were excluded. The relevant local feasibility committee approved the study protocol (reference 20231001). Of this cohort, electronic patient files of all patients were reviewed to identify the patients with a colovesical fistula. A colovesical fistula was defined as the presence of (1) at least one of three predominant symptoms (pneumaturia, fecaluria, or recurrent UTI) and (2) diagnostic imaging (computed tomography scan (CT scan), cystoscopy, or cystography) suspected for a colovesical fistula. All patients who were suspected of having colovesical fistula underwent CT scan without the usage of vesical contrast. According to the local protocol of our hospital, performance of cystography was advised before IUC removal 5–7 days after surgery. However, whether cystography was performed and the exact timing of cystography was determined at the discretion of the attending surgeon. The intraoperative use of methylene blue for identification of the vesical defect and the placement of pelvic drain were finally decided by the attending surgeon. Primary endpoints of the study were the presence/absence of extravesical contrast on postoperative cystography and its effect on IUC duration, hospital stay, and postoperative (urological) complications. As a secondary outcome, a comparative analysis was performed of surgical outcome for patients with and without a colovesical fistula. Baseline patient characteristics such as gender, age, body mass index (BMI), American Society of Anesthesiologists (ASA) classification, and previous abdominal surgery were obtained. Furthermore, surgical characteristics including approach, duration, fistula visualization, intraoperative bladder management, colonic anastomosis, and complications according to the Clavien–Dindo classification were obtained [ 21 ]. In addition, attention was directed towards the postoperative course. Variables such as length of hospital stay, IUC duration, urological complications, and fistula recurrence were collected. Baseline characteristics are displayed using mean and standard deviation in case of normal distribution or median and range in case of non-normal distribution. Discrete and categorical variables are presented as proportions of patients (percentages). Chi-square tests, Fisher’s exact tests, t  tests, and Mann–Whitney U  tests were used to investigate differences between demographic characteristics and to assess primary and secondary outcomes. Levene’s test for variance has been used for assessing equal variance between groups. Normality of distribution was analyzed using Shapiro–Wilk’s tests and normality plots. A value of p  < 0.05 was considered statistically significant. Data was analyzed using SPSS statistics (version 28, IBM Corporation, Armonk, New York, USA).