Surg Endosc (2014) 28:961–966 DOI 10.1007/s00464-013-3257-0
and Other Interventional Techniques
Risk of colon cancer after computed tomography-diagnosed acute diverticulitis: is routine colonoscopy necessary? Ville Sallinen • Panu Mentula • Ari Leppa¨niemi
Received: 27 June 2013 / Accepted: 30 September 2013 / Published online: 1 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Background Current guidelines recommend computed tomography (CT) for diagnosing diverticulitis and for routine follow-up colonoscopy to rule out cancer. Scientific data to support routine colonoscopy after acute diverticulitis are scarce and conflicting. This study aimed to evaluate the risk of colon cancer mimicking diverticulitis, and hence the need for routine colonoscopy after CT-diagnosed acute diverticulitis. Methods This study was a retrospective analysis of patients treated for acute diverticulitis in a single academic institution during 2006–2010. Data regarding age, sex, laboratory parameters, prior diverticulitis, surgical operations, pathology reports, and CT characteristics were collected. Risk factors for finding colon cancer after CTdiagnosed acute diverticulitis were identified by multivariate analysis. Results The study enrolled 633 patients with CT-diagnosed acute diverticulitis. Of these patients, 97 underwent emergency resection, whereas 536 were treated conservatively, 394 of whom underwent colonoscopy. The findings showed 17 cancers (2.7 %) in patients with an initial diagnosis of acute diverticulitis. As shown by CT, 16 cancer patients (94 %) had abscess, whereas one patient had pericolic extraluminal air but no abscess. Of the patients with abscess, 11.4 % had cancer mimicking acute diverticulitis. No cancer was found in the patients with
V. Sallinen (&) P. Mentula A. Leppa¨niemi Department of Abdominal Surgery, Helsinki University Central Hospital, Haartmaninkatu 4, 00029 HUS Helsinki, Finland e-mail: [email protected]
uncomplicated diverticulitis. Besides abscess, other independent risk factors for cancer included suspicion of cancer by a radiologist, thickness of the bowel wall exceeding 15 mm, no diverticula observed, and previously undiagnosed metastases. Conclusions Routine colonoscopy after CT-proven uncomplicated diverticulitis seems to be unnecessary, but colonoscopy should be performed for patients with a diagnosis of diverticular abscess. Keywords Diverticulosis Diverticular disease Sigmoidoscopy Colorectal cancer
Diverticulosis is a common disease affecting 20–60 % of the population. Patients with diverticulosis have *20–25 % lifetime risk of acute diverticulitis. Current guidelines recommend the use of ultrasound or computed tomography (CT) to diagnose acute diverticulitis and a subsequent colonoscopy to rule out cancer-mimicking diverticulitis [1–4]. Colonoscopy is a time-consuming and invasive method that carries a risk for complications. In recent cohorts of 20,000 and 57,000 patients, the risk for iatrogenic large bowel perforation was 0.04 and 0.07 %, respectively [5, 6]. The need for routine colonoscopy after acute diverticulitis has been questioned recently, with some authors deeming it unnecessary [7–9]. Nevertheless, *0.5–2 % of CT-diagnosed acute diverticulitis cases turn out to be colon cancers instead [8–11]. A recent metaanalysis of 10 articles reporting on a total of 771 patients concluded that current data supporting the recommendation of routine colonoscopy after acute diverticulitis are limited [12]. This study aimed to identify patient and CT characteristics associated with the risk for colon cancer in patients with CT-diagnosed acute diverticulitis. Identification of
123
962
risk factors will help in directing resources to patients most likely to benefit from colonoscopy. We report the findings and risk factors for cancer observed in 633 patients treated for CT-diagnosed acute diverticulitis.
Materials and methods This retrospective cohort study was conducted in an academic hospital center that serves as both a secondary (city hospital) and a tertiary referral center for a population of 1.5 million inhabitants. The study enrolled patients with International Classification of Diseases (ICD)-10 diagnosis code K57 (diverticular disease of the intestine) at any point during admission in 2006–2010. The study was approved by the Hospital Ethics Committee, the City Research Committee, and the National Institute for Health and Welfare. Patient data were obtained from electronic patient records and manually analyzed for age, sex, laboratory parameters, prior diverticulitis, surgical operations, and pathology reports. To obtain exact measurements, CT reports were obtained and analyzed, and CT images were reexamined. The CT reports and images were analyzed with blinding to other patient characteristics. At our institution, on-call radiologists, who can be either consultants or residents, generate initial CT reports. An expert consultant radiologist specialized in abdominal radiology provides a complementary report. The site of the affected colon segment and the presence of abscess, ascites, extraluminal air, colon occlusion, diverticula, lymphadenopathy, ascites, or distant metastases were retrieved from the final complementary CT report. The thickness and length of the affected colon wall segment were measured from CT images. In this study, suspicion of cancer by a radiologist was defined as any reference to possible cancer in the radiology report including, but not limited to, the following phrases: ‘‘cancer cannot be ruled out,’’ ‘‘raises concern of cancer,’’ and ‘‘endoscopic evaluation should be carried out to confirm diagnosis.’’ A CT diagnosis of diverticulitis was defined as an acute diverticulitis given as a primary diagnosis by an expert consultant radiologist in the complementary report. A CT-diagnosed uncomplicated diverticulitis was defined as an acute diverticulitis without extraluminal air, abscess, stricture, or fistula. Colonoscopy charts were retrieved from electronic patient records of all district hospitals as well as city hospitals. These sources include all the publicly produced colonoscopies within the Helsinki area. Only colonoscopies performed at or after admission for the index diverticulitis were included in the study. All available electronic patient records of the hospital district were searched manually for colorectal cancer in patients not evaluated by
123
Surg Endosc (2014) 28:961–966
colonoscopy and those evaluated by incomplete colonoscopy (no cecal intubation or identification of the valvula). It should be noted that surgical treatment of cancer in private hospitals of Finland is rare. Nearly all cases are treated in public hospitals. Furthermore, because it is mandatory by Finnish law to report all cancer cases, a similar search was performed in the Finnish Cancer Registry (Institute of Statistical and Epidemiological Cancer Research, Helsinki, Finland) with a minimum 2-year interval after an episode of acute diverticulitis. Statistical analyses, including chi-square and logistic regression analysis with forward selection, were performed using SPSS Statistics 21 (IBM, Armonk, NY, USA).
Results The ICD-10 code search identified 1,236 cases, and 603 of these cases were primarily excluded (Fig. 1). Altogether, 633 patients living within our catchment area were identified as having a clinically and CT-diagnosed acute diverticulitis of the large intestine. The mean age of the patients was 58.5 ± 13.9 years, and 62.2 % of the patients were women. Intravenous contrast medium was used for the CT in 549 cases (87 %) and oral contrast medium in 574 cases (91 %). Rectal contrast medium was not used in any patient. The use of CT to diagnose acute diverticulitis increased from 49 % in 2006 to 81 % in 2010. Whereas 475 of the patients had experienced their first attack of acute diverticulitis, 158 had experienced recurrent diverticulitis. The cases comprised 604 patients with left-sided diverticulitis, 9 with transverse colon diverticulitis, and 20 with right-sided diverticulitis. Of these cases, 438 had no extraluminal air, 110 had pericolic air, 18 had distant air in the retroperitoneum, and 75 had distant air in the abdominal cavity. There were 145 cases of intraabdominal abscess and 2 cases of colovesical fistula. Among the patients, 125 had one abscess, 15 had two abscesses, and 5 had three abscesses. The mean size of the largest abscess was 51 ± 32 mm. Multivariate analysis recognized abscess, suspicion of cancer by a radiologist, a colonic wall thickness of 15 mm or greater, no diverticula, and previously undiagnosed metastases as independent risk factors for diverticulitis to be cancer instead (Table 1). None of these risk factors were found in 349 of the patients (55 %). For 97 patients treated as having an acute diverticulitis, emergency surgery was performed during the index admission. Cancer was found in seven patients either during the surgery or in the pathology report of the resected specimen. Of the 536 remaining patients treated conservatively, 394 underwent endoscopic colonic evaluation.
Surg Endosc (2014) 28:961–966
963
Fig. 1 Flow chart showing patient selection and results. Table 2 presents specific reasons for omitting colonoscopy. pt patient
For the reasons listed in Table 2, 142 patients did not undergo colonoscopy. The mean interval from discharge to colonoscopy was 122 ± 180 days. Nine adenocarcinomas of the colon and one appendix mucocele were found via colonoscopy. The mean interval from discharge to the colonoscopy that found a cancer was 55 ± 45 days (range 6–141 days). The adenocarcinomas were located at the site of the affected colon segment presumed to contain acute diverticulitis. Appendiceal mucocele was found in a patient with sigmoid diverticulitis. The mucocele was located adjacent to this affected colon segment, and the radiology report stated that it represented three abscesses associated with diverticulitis. Of the 17 cancers, 16 were found in patients with abscess, and 1 cancer was found in a patient with pericolic air. In other
words, 16 of 145 abscess cases were found to harbor cancer, yielding an 11.4 % risk of cancer among cases in which abscess was seen by CT. No cancers were found in patients with a CT-diagnosed uncomplicated acute diverticulitis. Of 384 cancer-negative endoscopies, 85 were subtotal colonoscopies, including sigmoidoscopies. No further colorectal cancers were found in patients with an incomplete (n = 85) or no (n = 142) colonic evaluation in the search of hospital medical records or the Finnish Cancer Registry (Fig. 1). Endoscopic evaluation of the colon showed 45 patients with a polyp or polyps and 49 patients with hyperplastic polyps (Table 3). Excluding the cancer cases, diverticulosis was found in 357 patients (93 %). Additionally, one World Health Organization (WHO) grade 1 neuroendocrine tumor (NET) was found in the rectum of a patient treated for
123
964
Surg Endosc (2014) 28:961–966
Table 1 Univariate chi-square and multivariate logistic regression analysis with forward selection of risk factors for diverticulitis-mimicking cancer in patients with CT-diagnosed acute diverticulitis
Age [60 years
Cancer found (n = 17) n (%)
Acute diverticulitis, no cancer (n = 616) n (%)
Univariate chi-square OR (95 % CI)
Multivariate logistic regression with forward stepwise selection OR (95 % CI)
12 (71)
253 (41)
3.4 (1.2–9.9)
NS
Recurrent diverticulitis
3 (18)
155 (25)
0.6 (0.2–2.2)
NS
Anemiaa
8 (47)
101 (16)
4.5 (1.7–12.0)
NS
Abscess
16 (94)
129 (21)
60.4 (7.9–459.7)
Extraluminal air without abscess
1 (6)
148 (24)
0.2 (0.03–1.5)
NS
Ascites
2 (12)
90 (15)
0.8 (0.2–3.5)
NS
21.2 (7.7–58.8)
Suspicion of cancer by radiologist
40.7 (4.8–344)
10 (59)
31 (5)
Signs of colon occlusion
0 (0)
17 (3)
Affected colon shorter than 10 cm
9 (53)
272 (44)
1.4 (0.5–3.7)
NS
Thickness of affected colon C15 mm Affected colon segment other than sigmoid colon
9 (53) 2 (12)
130 (21) 104 (17)
4.2 (1.6–11.1) 0.6 (0.1–2.9)
17.5 (3.3–91) NS
No diverticulosis of the affected colon segment
1 (1–1)
10 (59)
34 (6)
23.7 (8.5–66.1)
Mesenterial or paraaortic lymphadenopathy
5 (29)
11 (18)
22.9 (6.8–76.2)
Undiagnosed metastases
3 (18)
1 (0.2)
131.8 (12.9–1346.8)
7.7 (1.9–31.4) NS
16.1 (3.2–79.0) NS 36.9 (1.2–1151.3)
OR odds ratio, CI confidence interval, NS not selected in forward analysis a
Anemia is defined as a hemoglobin concentration level below 11.7 g/dL in women and below 13.4 g/dL in men, which are Finnish national reference values
Table 2 Reasons for no colonoscopy Reason for no colonoscopy
Table 3 Incidental findings in colonoscopy n = 142
Colonoscopy within 2 years
40
Colonoscopy earlier, but more than 2 years ago
22
Barium enema within 2 years Barium enema earlier, but more than 2 years ago Patient declined
Incidental findings in colonoscopy
(n = 384) n (%)
No incidental findings
289 (75)
1
Polyp(s), mild dysplasia
1
Polyp(s), moderate dysplasia
23
Polyp(s), severe dysplasia
Patient died
2
Polyp(s), hyperplastic
Colonography ordered instead
3
Rectal NET, grade 1
3
NET neuroendocrine tumor
Elective sigmoid resection before colonoscopy Patient too old/frail Colonoscopy at a private endoscopist, and charts are thus unavailable Not scheduled, reason unknown Colonoscopy scheduled, but colonoscopy charts not found
1 (0.3) 4 (1) 49 (13) 1 (0.3)
18 2 9 18
uncomplicated diverticulitis of the descending colon. This NET was considered incidental and treated by polypectomy, and thus was not included in the analysis of cancers. No cases of colitis ulcerosa or Crohn’s disease were found.
Discussion Perforated colon cancer with abscess is not uncommon among patients with a diagnosis of diverticular abscess. Nearly all the cases (16 of 17 cancers, 94 %) involved an
123
40 (10)
abscess in which a CT-diagnosed acute diverticulitis turned out to be cancer. Other independent risk factors included suspicion of cancer by a radiologist, colonic wall thickening greater than 15 mm, no diverticula observed, and previously undiagnosed distant metastases. A recent metaanalysis evaluated 10 articles reporting on a total of 771 patients and concluded that scientific data supporting routine colonoscopy after acute diverticulitis are limited [12]. However, recent studies have provided more data. In 2011, Westwood et al. [9] evaluated 292 patients with acute uncomplicated diverticulitis diagnosed by strict CT criteria. Only one colon cancer case was found, and the authors concluded that colonoscopy is not needed in the absence of other indications. In 2012, Schmilovitz-Weiss et al. [7] evaluated colonoscopies of 100 patients treated for acute diverticulitis
Surg Endosc (2014) 28:961–966
(uncomplicated and complicated). They found no colorectal cancers shown by follow-up colonoscopy and concluded that colonoscopy does not affect the management of patients treated for acute diverticulitis. In 2012, Schout et al. [8] studied 516 patients with acute diverticulitis, 423 of whom were endoscopically or radiologically evaluated in the follow-up assessment. They found eight patients with colorectal cancer, six of whom had alarming symptoms such as rectal bleeding, persisting pain, or weight loss. They concluded that colonoscopy is indicated after conservatively treated acute diverticulitis only for symptomatic patients. However, another study by Lau et al. [11] in 2011 recommended routine colonoscopy. In their study, 319 followup colonoscopies were performed in 1,088 cases of acute diverticulitis, with the remaining cases detected in the Australian Cancer Registry. Altogether, 23 cancers were found, and local perforation, abscess, and fistula were recognized as risk factors for a colorectal cancer diagnosis. Our findings are in line with these results, and abscess seems to be a strong risk factor for cancer. Our findings showed only two fistulas, neither of which was malignant. The low count of fistulas can be explained by the fact that the fistulas were coded using a different ICD-10 diagnosis code (N32.1 for vesicointestinal fistula, N82.3 for fistula of the vagina to the large intestine) and thus were not included in the initial search. Our findings showed no cancers detected in patients with CT-diagnosed uncomplicated acute diverticulitis. One cancer emerged in a case of acute diverticulitis with pericolic air. All the other cancer cases were found in patients with abscess. Among the CT-diagnosed acute diverticulitis cases with abscess, 11.4 % turned out to be cancer cases. In our opinion, patients who have complicated diverticulitis with abscess clearly warrant endoscopic evaluation of the colon. Another indication for colonoscopy is the detection of adenomas. Our findings showed only four adenomas with severe dysplasia and one adenoma with moderate dysplasia. Also, a few reports indicated that diverticular disease or diverticulitis per se does not increase the risk for colonic polyps [13, 14]. Thus, it seems that uncomplicated diverticilitis is not a good indication for routine colonoscopy. However, it must be borne in mind that uncomplicated diverticulitis should not hinder colonoscopy if other indications, including colorectal cancer screening, are present. The current guidelines recommend colonoscopy 6 weeks after the acute phase [1, 2]. Studies have shown that colonoscopy can be performed in uncomplicated cases early, even during the same admission [15]. However, early colonoscopy is not advised for patients with complicated disease due to fear of perforation [15]. Only one pilot study reported a perforation of sigmoid colon in one patient with pericolic air when colonoscopy was performed at the index
965
admission [10]. Thus, the scientific basis for not recommending early colonoscopy in cases of complicated diverticulitis is limited, and optimal timing of colonoscopy after complicated diverticulitis is not known. At our institution, colonoscopy is performed 4–6 weeks after acute complicated diverticulitis, whereas CT-colonoscopy is reserved for patients whose colonoscopy has failed or is contraindicated. Our study had some limitations. First, only patients with aptly coded diverticulitis were included in the study. There may have been patients with acute diverticulitis who had different diagnosis codes such as C18-20 for suspected colorectal cancer. In any case, these patients would undergo colonoscopy because suspicion of cancer had been raised. Second, the diagnosis of acute colonic diverticulitis was based on the CT report given by an expert abdominal consultant radiologist. Despite this, in 44 cases of acute diverticulitis, no diverticula were seen. It could therefore be argued that these cases had an incorrect diagnosis in the first place. However, although 10 of these cases turned out to be cancers instead, we note that 26 of the remaining 34 patients underwent an endoscopy (n = 13) or emergency operation (n = 13) that confirmed the diagnosis to be diverticulosis of the affected segment. It is thus obvious that strict CT criteria including visible diverticula do not yield 100 % sensitivity. Because the findings showed that a lack of diverticulosis in the CT scan independently increased the risk of discovering colon cancer to 16-fold in this study, this situation should always lead to colonoscopic confirmation of the diagnosis. In conclusion, routine colonoscopy should be performed after nonoperatively treated diverticular abscess to rule out colon cancer in the affected part of the colon. Also, patients undergoing operative treatment for diverticular abscess should have an endoscopy before the operation to exclude perforation of colon cancer. Routine colonoscopy after uncomplicated acute diverticulitis seems unnecessary in the absence of other risk factors. Acknowledgments We thank Dr. Ari Aimolahti from the City of Helsinki for help in retrieving the colonoscopy charts and Professor Nea Malila for providing data from the Finnish Cancer Registry. The Martti I. Turunen foundation provided support for this study. Disclosures Ville Sallinen, Panu Mentula, and Ari Leppa¨niemi have no conflicts of interest or financial ties to disclose.
References 1. Fozard JBJ, Armitage NC, Schofield JB, Jones OM (2011) ACPGBI position statement on elective resection for diverticulitis. Colorectal Dis 13:1–11. doi:10.1111/j.1463-1318.2010.02531.x
123
966 2. Andersen JC, Bundgaard L, Elbrønd H et al (2012) Danish national guidelines for treatment of diverticular disease. Dan Med J 59:C4453 3. Rafferty J, Shellito P, Hyman NH, Buie WD (2006) Practice parameters for sigmoid diverticulitis. Dis Colon Rectum 49:939–944. doi:10.1007/s10350-006-0578-2 4. Jacobs DO (2007) Clinical practice: diverticulitis. N Engl J Med 357:2057–2066. doi:10.1056/NEJMcp073228 5. Gavin DR, Valori RM, Anderson JT et al (2013) The national colonoscopy audit: a nationwide assessment of the quality and safety of colonoscopy in the UK. Gut 62:242–249. doi:10.1136/ gutjnl-2011-301848 6. Samalavicius NE, Kazanavicius D, Lunevicius R et al (2012) Incidence, risk, management, and outcomes of iatrogenic fullthickness large bowel injury associated with 56,882 colonoscopies in 14 Lithuanian hospitals. Surg Endosc 27(5):1628–1635. doi:10.1007/s00464-012-2642-4 7. Schmilovitz-Weiss H, Yalunin E, Boaz M et al (2012) Does a colonoscopy after acute diverticulitis affect its management? A single-center experience. J Clin Gastroenterol 46:317–320. doi:10.1097/MCG.0b013e31823a43a1 8. Schout PJC, Spillenaar Bilgen EJ, Groenen MJM (2012) Routine screening for colon cancer after conservative treatment of diverticulitis. Dig Surg 29:408–411. doi:10.1159/000345332 9. Westwood DA, Eglinton TW, Frizelle FA (2011) Routine colonoscopy following acute uncomplicated diverticulitis. Br J Surg 98:1630–1634. doi:10.1002/bjs.7602
123
Surg Endosc (2014) 28:961–966 10. Sakhnini E, Lahat A, Melzer E et al (2004) Early colonoscopy in patients with acute diverticulitis: results of a prospective pilot study. Endoscopy 36:504–507. doi:10.1055/s-2004-814398 11. Lau KC, Spilsbury K, Farooque Y et al (2011) Is colonoscopy still mandatory after a CT diagnosis of left-sided diverticulitis: can colorectal cancer be confidently excluded? Dis Colon Rectum 54:1265–1270. doi:10.1097/DCR.0b013e31822899a2 12. Sai VF, Velayos F, Neuhaus J, Westphalen AC (2012) Colonoscopy after CT diagnosis of diverticulitis to exclude colon cancer: a systematic literature review. Radiology 263:383–390. doi:10.1148/radiol.12111869 13. Meurs-Szojda MM, Droste JSTS, Kuik DJ et al (2008) Diverticulosis and diverticulitis form no risk for polyps and colorectal neoplasia in 4,241 colonoscopies. Int J Colorectal Dis 23:979–984. doi:10.1007/s00384-008-0510-4 14. Lam TJ, Meurs-Szojda MM, Gundlach L et al (2009) There is no increased risk for colorectal cancer and adenomas in patients with diverticulitis: a retrospective longitudinal study. Colorectal Dis 12:1122–1126. doi:10.1111/j.1463-1318.2009.01992.x 15. Lahat A, Yanai H, Menachem Y et al (2007) The feasibility and risk of early colonoscopy in acute diverticulitis: a prospective controlled study. Endoscopy 39:521–524. doi:10.1055/s-2007-966399
and Other Interventional Techniques
Risk of colon cancer after computed tomography-diagnosed acute diverticulitis: is routine colonoscopy necessary? Ville Sallinen • Panu Mentula • Ari Leppa¨niemi
Received: 27 June 2013 / Accepted: 30 September 2013 / Published online: 1 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Background Current guidelines recommend computed tomography (CT) for diagnosing diverticulitis and for routine follow-up colonoscopy to rule out cancer. Scientific data to support routine colonoscopy after acute diverticulitis are scarce and conflicting. This study aimed to evaluate the risk of colon cancer mimicking diverticulitis, and hence the need for routine colonoscopy after CT-diagnosed acute diverticulitis. Methods This study was a retrospective analysis of patients treated for acute diverticulitis in a single academic institution during 2006–2010. Data regarding age, sex, laboratory parameters, prior diverticulitis, surgical operations, pathology reports, and CT characteristics were collected. Risk factors for finding colon cancer after CTdiagnosed acute diverticulitis were identified by multivariate analysis. Results The study enrolled 633 patients with CT-diagnosed acute diverticulitis. Of these patients, 97 underwent emergency resection, whereas 536 were treated conservatively, 394 of whom underwent colonoscopy. The findings showed 17 cancers (2.7 %) in patients with an initial diagnosis of acute diverticulitis. As shown by CT, 16 cancer patients (94 %) had abscess, whereas one patient had pericolic extraluminal air but no abscess. Of the patients with abscess, 11.4 % had cancer mimicking acute diverticulitis. No cancer was found in the patients with
V. Sallinen (&) P. Mentula A. Leppa¨niemi Department of Abdominal Surgery, Helsinki University Central Hospital, Haartmaninkatu 4, 00029 HUS Helsinki, Finland e-mail: [email protected]
uncomplicated diverticulitis. Besides abscess, other independent risk factors for cancer included suspicion of cancer by a radiologist, thickness of the bowel wall exceeding 15 mm, no diverticula observed, and previously undiagnosed metastases. Conclusions Routine colonoscopy after CT-proven uncomplicated diverticulitis seems to be unnecessary, but colonoscopy should be performed for patients with a diagnosis of diverticular abscess. Keywords Diverticulosis Diverticular disease Sigmoidoscopy Colorectal cancer
Diverticulosis is a common disease affecting 20–60 % of the population. Patients with diverticulosis have *20–25 % lifetime risk of acute diverticulitis. Current guidelines recommend the use of ultrasound or computed tomography (CT) to diagnose acute diverticulitis and a subsequent colonoscopy to rule out cancer-mimicking diverticulitis [1–4]. Colonoscopy is a time-consuming and invasive method that carries a risk for complications. In recent cohorts of 20,000 and 57,000 patients, the risk for iatrogenic large bowel perforation was 0.04 and 0.07 %, respectively [5, 6]. The need for routine colonoscopy after acute diverticulitis has been questioned recently, with some authors deeming it unnecessary [7–9]. Nevertheless, *0.5–2 % of CT-diagnosed acute diverticulitis cases turn out to be colon cancers instead [8–11]. A recent metaanalysis of 10 articles reporting on a total of 771 patients concluded that current data supporting the recommendation of routine colonoscopy after acute diverticulitis are limited [12]. This study aimed to identify patient and CT characteristics associated with the risk for colon cancer in patients with CT-diagnosed acute diverticulitis. Identification of
123
962
risk factors will help in directing resources to patients most likely to benefit from colonoscopy. We report the findings and risk factors for cancer observed in 633 patients treated for CT-diagnosed acute diverticulitis.
Materials and methods This retrospective cohort study was conducted in an academic hospital center that serves as both a secondary (city hospital) and a tertiary referral center for a population of 1.5 million inhabitants. The study enrolled patients with International Classification of Diseases (ICD)-10 diagnosis code K57 (diverticular disease of the intestine) at any point during admission in 2006–2010. The study was approved by the Hospital Ethics Committee, the City Research Committee, and the National Institute for Health and Welfare. Patient data were obtained from electronic patient records and manually analyzed for age, sex, laboratory parameters, prior diverticulitis, surgical operations, and pathology reports. To obtain exact measurements, CT reports were obtained and analyzed, and CT images were reexamined. The CT reports and images were analyzed with blinding to other patient characteristics. At our institution, on-call radiologists, who can be either consultants or residents, generate initial CT reports. An expert consultant radiologist specialized in abdominal radiology provides a complementary report. The site of the affected colon segment and the presence of abscess, ascites, extraluminal air, colon occlusion, diverticula, lymphadenopathy, ascites, or distant metastases were retrieved from the final complementary CT report. The thickness and length of the affected colon wall segment were measured from CT images. In this study, suspicion of cancer by a radiologist was defined as any reference to possible cancer in the radiology report including, but not limited to, the following phrases: ‘‘cancer cannot be ruled out,’’ ‘‘raises concern of cancer,’’ and ‘‘endoscopic evaluation should be carried out to confirm diagnosis.’’ A CT diagnosis of diverticulitis was defined as an acute diverticulitis given as a primary diagnosis by an expert consultant radiologist in the complementary report. A CT-diagnosed uncomplicated diverticulitis was defined as an acute diverticulitis without extraluminal air, abscess, stricture, or fistula. Colonoscopy charts were retrieved from electronic patient records of all district hospitals as well as city hospitals. These sources include all the publicly produced colonoscopies within the Helsinki area. Only colonoscopies performed at or after admission for the index diverticulitis were included in the study. All available electronic patient records of the hospital district were searched manually for colorectal cancer in patients not evaluated by
123
Surg Endosc (2014) 28:961–966
colonoscopy and those evaluated by incomplete colonoscopy (no cecal intubation or identification of the valvula). It should be noted that surgical treatment of cancer in private hospitals of Finland is rare. Nearly all cases are treated in public hospitals. Furthermore, because it is mandatory by Finnish law to report all cancer cases, a similar search was performed in the Finnish Cancer Registry (Institute of Statistical and Epidemiological Cancer Research, Helsinki, Finland) with a minimum 2-year interval after an episode of acute diverticulitis. Statistical analyses, including chi-square and logistic regression analysis with forward selection, were performed using SPSS Statistics 21 (IBM, Armonk, NY, USA).
Results The ICD-10 code search identified 1,236 cases, and 603 of these cases were primarily excluded (Fig. 1). Altogether, 633 patients living within our catchment area were identified as having a clinically and CT-diagnosed acute diverticulitis of the large intestine. The mean age of the patients was 58.5 ± 13.9 years, and 62.2 % of the patients were women. Intravenous contrast medium was used for the CT in 549 cases (87 %) and oral contrast medium in 574 cases (91 %). Rectal contrast medium was not used in any patient. The use of CT to diagnose acute diverticulitis increased from 49 % in 2006 to 81 % in 2010. Whereas 475 of the patients had experienced their first attack of acute diverticulitis, 158 had experienced recurrent diverticulitis. The cases comprised 604 patients with left-sided diverticulitis, 9 with transverse colon diverticulitis, and 20 with right-sided diverticulitis. Of these cases, 438 had no extraluminal air, 110 had pericolic air, 18 had distant air in the retroperitoneum, and 75 had distant air in the abdominal cavity. There were 145 cases of intraabdominal abscess and 2 cases of colovesical fistula. Among the patients, 125 had one abscess, 15 had two abscesses, and 5 had three abscesses. The mean size of the largest abscess was 51 ± 32 mm. Multivariate analysis recognized abscess, suspicion of cancer by a radiologist, a colonic wall thickness of 15 mm or greater, no diverticula, and previously undiagnosed metastases as independent risk factors for diverticulitis to be cancer instead (Table 1). None of these risk factors were found in 349 of the patients (55 %). For 97 patients treated as having an acute diverticulitis, emergency surgery was performed during the index admission. Cancer was found in seven patients either during the surgery or in the pathology report of the resected specimen. Of the 536 remaining patients treated conservatively, 394 underwent endoscopic colonic evaluation.
Surg Endosc (2014) 28:961–966
963
Fig. 1 Flow chart showing patient selection and results. Table 2 presents specific reasons for omitting colonoscopy. pt patient
For the reasons listed in Table 2, 142 patients did not undergo colonoscopy. The mean interval from discharge to colonoscopy was 122 ± 180 days. Nine adenocarcinomas of the colon and one appendix mucocele were found via colonoscopy. The mean interval from discharge to the colonoscopy that found a cancer was 55 ± 45 days (range 6–141 days). The adenocarcinomas were located at the site of the affected colon segment presumed to contain acute diverticulitis. Appendiceal mucocele was found in a patient with sigmoid diverticulitis. The mucocele was located adjacent to this affected colon segment, and the radiology report stated that it represented three abscesses associated with diverticulitis. Of the 17 cancers, 16 were found in patients with abscess, and 1 cancer was found in a patient with pericolic air. In other
words, 16 of 145 abscess cases were found to harbor cancer, yielding an 11.4 % risk of cancer among cases in which abscess was seen by CT. No cancers were found in patients with a CT-diagnosed uncomplicated acute diverticulitis. Of 384 cancer-negative endoscopies, 85 were subtotal colonoscopies, including sigmoidoscopies. No further colorectal cancers were found in patients with an incomplete (n = 85) or no (n = 142) colonic evaluation in the search of hospital medical records or the Finnish Cancer Registry (Fig. 1). Endoscopic evaluation of the colon showed 45 patients with a polyp or polyps and 49 patients with hyperplastic polyps (Table 3). Excluding the cancer cases, diverticulosis was found in 357 patients (93 %). Additionally, one World Health Organization (WHO) grade 1 neuroendocrine tumor (NET) was found in the rectum of a patient treated for
123
964
Surg Endosc (2014) 28:961–966
Table 1 Univariate chi-square and multivariate logistic regression analysis with forward selection of risk factors for diverticulitis-mimicking cancer in patients with CT-diagnosed acute diverticulitis
Age [60 years
Cancer found (n = 17) n (%)
Acute diverticulitis, no cancer (n = 616) n (%)
Univariate chi-square OR (95 % CI)
Multivariate logistic regression with forward stepwise selection OR (95 % CI)
12 (71)
253 (41)
3.4 (1.2–9.9)
NS
Recurrent diverticulitis
3 (18)
155 (25)
0.6 (0.2–2.2)
NS
Anemiaa
8 (47)
101 (16)
4.5 (1.7–12.0)
NS
Abscess
16 (94)
129 (21)
60.4 (7.9–459.7)
Extraluminal air without abscess
1 (6)
148 (24)
0.2 (0.03–1.5)
NS
Ascites
2 (12)
90 (15)
0.8 (0.2–3.5)
NS
21.2 (7.7–58.8)
Suspicion of cancer by radiologist
40.7 (4.8–344)
10 (59)
31 (5)
Signs of colon occlusion
0 (0)
17 (3)
Affected colon shorter than 10 cm
9 (53)
272 (44)
1.4 (0.5–3.7)
NS
Thickness of affected colon C15 mm Affected colon segment other than sigmoid colon
9 (53) 2 (12)
130 (21) 104 (17)
4.2 (1.6–11.1) 0.6 (0.1–2.9)
17.5 (3.3–91) NS
No diverticulosis of the affected colon segment
1 (1–1)
10 (59)
34 (6)
23.7 (8.5–66.1)
Mesenterial or paraaortic lymphadenopathy
5 (29)
11 (18)
22.9 (6.8–76.2)
Undiagnosed metastases
3 (18)
1 (0.2)
131.8 (12.9–1346.8)
7.7 (1.9–31.4) NS
16.1 (3.2–79.0) NS 36.9 (1.2–1151.3)
OR odds ratio, CI confidence interval, NS not selected in forward analysis a
Anemia is defined as a hemoglobin concentration level below 11.7 g/dL in women and below 13.4 g/dL in men, which are Finnish national reference values
Table 2 Reasons for no colonoscopy Reason for no colonoscopy
Table 3 Incidental findings in colonoscopy n = 142
Colonoscopy within 2 years
40
Colonoscopy earlier, but more than 2 years ago
22
Barium enema within 2 years Barium enema earlier, but more than 2 years ago Patient declined
Incidental findings in colonoscopy
(n = 384) n (%)
No incidental findings
289 (75)
1
Polyp(s), mild dysplasia
1
Polyp(s), moderate dysplasia
23
Polyp(s), severe dysplasia
Patient died
2
Polyp(s), hyperplastic
Colonography ordered instead
3
Rectal NET, grade 1
3
NET neuroendocrine tumor
Elective sigmoid resection before colonoscopy Patient too old/frail Colonoscopy at a private endoscopist, and charts are thus unavailable Not scheduled, reason unknown Colonoscopy scheduled, but colonoscopy charts not found
1 (0.3) 4 (1) 49 (13) 1 (0.3)
18 2 9 18
uncomplicated diverticulitis of the descending colon. This NET was considered incidental and treated by polypectomy, and thus was not included in the analysis of cancers. No cases of colitis ulcerosa or Crohn’s disease were found.
Discussion Perforated colon cancer with abscess is not uncommon among patients with a diagnosis of diverticular abscess. Nearly all the cases (16 of 17 cancers, 94 %) involved an
123
40 (10)
abscess in which a CT-diagnosed acute diverticulitis turned out to be cancer. Other independent risk factors included suspicion of cancer by a radiologist, colonic wall thickening greater than 15 mm, no diverticula observed, and previously undiagnosed distant metastases. A recent metaanalysis evaluated 10 articles reporting on a total of 771 patients and concluded that scientific data supporting routine colonoscopy after acute diverticulitis are limited [12]. However, recent studies have provided more data. In 2011, Westwood et al. [9] evaluated 292 patients with acute uncomplicated diverticulitis diagnosed by strict CT criteria. Only one colon cancer case was found, and the authors concluded that colonoscopy is not needed in the absence of other indications. In 2012, Schmilovitz-Weiss et al. [7] evaluated colonoscopies of 100 patients treated for acute diverticulitis
Surg Endosc (2014) 28:961–966
(uncomplicated and complicated). They found no colorectal cancers shown by follow-up colonoscopy and concluded that colonoscopy does not affect the management of patients treated for acute diverticulitis. In 2012, Schout et al. [8] studied 516 patients with acute diverticulitis, 423 of whom were endoscopically or radiologically evaluated in the follow-up assessment. They found eight patients with colorectal cancer, six of whom had alarming symptoms such as rectal bleeding, persisting pain, or weight loss. They concluded that colonoscopy is indicated after conservatively treated acute diverticulitis only for symptomatic patients. However, another study by Lau et al. [11] in 2011 recommended routine colonoscopy. In their study, 319 followup colonoscopies were performed in 1,088 cases of acute diverticulitis, with the remaining cases detected in the Australian Cancer Registry. Altogether, 23 cancers were found, and local perforation, abscess, and fistula were recognized as risk factors for a colorectal cancer diagnosis. Our findings are in line with these results, and abscess seems to be a strong risk factor for cancer. Our findings showed only two fistulas, neither of which was malignant. The low count of fistulas can be explained by the fact that the fistulas were coded using a different ICD-10 diagnosis code (N32.1 for vesicointestinal fistula, N82.3 for fistula of the vagina to the large intestine) and thus were not included in the initial search. Our findings showed no cancers detected in patients with CT-diagnosed uncomplicated acute diverticulitis. One cancer emerged in a case of acute diverticulitis with pericolic air. All the other cancer cases were found in patients with abscess. Among the CT-diagnosed acute diverticulitis cases with abscess, 11.4 % turned out to be cancer cases. In our opinion, patients who have complicated diverticulitis with abscess clearly warrant endoscopic evaluation of the colon. Another indication for colonoscopy is the detection of adenomas. Our findings showed only four adenomas with severe dysplasia and one adenoma with moderate dysplasia. Also, a few reports indicated that diverticular disease or diverticulitis per se does not increase the risk for colonic polyps [13, 14]. Thus, it seems that uncomplicated diverticilitis is not a good indication for routine colonoscopy. However, it must be borne in mind that uncomplicated diverticulitis should not hinder colonoscopy if other indications, including colorectal cancer screening, are present. The current guidelines recommend colonoscopy 6 weeks after the acute phase [1, 2]. Studies have shown that colonoscopy can be performed in uncomplicated cases early, even during the same admission [15]. However, early colonoscopy is not advised for patients with complicated disease due to fear of perforation [15]. Only one pilot study reported a perforation of sigmoid colon in one patient with pericolic air when colonoscopy was performed at the index
965
admission [10]. Thus, the scientific basis for not recommending early colonoscopy in cases of complicated diverticulitis is limited, and optimal timing of colonoscopy after complicated diverticulitis is not known. At our institution, colonoscopy is performed 4–6 weeks after acute complicated diverticulitis, whereas CT-colonoscopy is reserved for patients whose colonoscopy has failed or is contraindicated. Our study had some limitations. First, only patients with aptly coded diverticulitis were included in the study. There may have been patients with acute diverticulitis who had different diagnosis codes such as C18-20 for suspected colorectal cancer. In any case, these patients would undergo colonoscopy because suspicion of cancer had been raised. Second, the diagnosis of acute colonic diverticulitis was based on the CT report given by an expert abdominal consultant radiologist. Despite this, in 44 cases of acute diverticulitis, no diverticula were seen. It could therefore be argued that these cases had an incorrect diagnosis in the first place. However, although 10 of these cases turned out to be cancers instead, we note that 26 of the remaining 34 patients underwent an endoscopy (n = 13) or emergency operation (n = 13) that confirmed the diagnosis to be diverticulosis of the affected segment. It is thus obvious that strict CT criteria including visible diverticula do not yield 100 % sensitivity. Because the findings showed that a lack of diverticulosis in the CT scan independently increased the risk of discovering colon cancer to 16-fold in this study, this situation should always lead to colonoscopic confirmation of the diagnosis. In conclusion, routine colonoscopy should be performed after nonoperatively treated diverticular abscess to rule out colon cancer in the affected part of the colon. Also, patients undergoing operative treatment for diverticular abscess should have an endoscopy before the operation to exclude perforation of colon cancer. Routine colonoscopy after uncomplicated acute diverticulitis seems unnecessary in the absence of other risk factors. Acknowledgments We thank Dr. Ari Aimolahti from the City of Helsinki for help in retrieving the colonoscopy charts and Professor Nea Malila for providing data from the Finnish Cancer Registry. The Martti I. Turunen foundation provided support for this study. Disclosures Ville Sallinen, Panu Mentula, and Ari Leppa¨niemi have no conflicts of interest or financial ties to disclose.
References 1. Fozard JBJ, Armitage NC, Schofield JB, Jones OM (2011) ACPGBI position statement on elective resection for diverticulitis. Colorectal Dis 13:1–11. doi:10.1111/j.1463-1318.2010.02531.x
123
966 2. Andersen JC, Bundgaard L, Elbrønd H et al (2012) Danish national guidelines for treatment of diverticular disease. Dan Med J 59:C4453 3. Rafferty J, Shellito P, Hyman NH, Buie WD (2006) Practice parameters for sigmoid diverticulitis. Dis Colon Rectum 49:939–944. doi:10.1007/s10350-006-0578-2 4. Jacobs DO (2007) Clinical practice: diverticulitis. N Engl J Med 357:2057–2066. doi:10.1056/NEJMcp073228 5. Gavin DR, Valori RM, Anderson JT et al (2013) The national colonoscopy audit: a nationwide assessment of the quality and safety of colonoscopy in the UK. Gut 62:242–249. doi:10.1136/ gutjnl-2011-301848 6. Samalavicius NE, Kazanavicius D, Lunevicius R et al (2012) Incidence, risk, management, and outcomes of iatrogenic fullthickness large bowel injury associated with 56,882 colonoscopies in 14 Lithuanian hospitals. Surg Endosc 27(5):1628–1635. doi:10.1007/s00464-012-2642-4 7. Schmilovitz-Weiss H, Yalunin E, Boaz M et al (2012) Does a colonoscopy after acute diverticulitis affect its management? A single-center experience. J Clin Gastroenterol 46:317–320. doi:10.1097/MCG.0b013e31823a43a1 8. Schout PJC, Spillenaar Bilgen EJ, Groenen MJM (2012) Routine screening for colon cancer after conservative treatment of diverticulitis. Dig Surg 29:408–411. doi:10.1159/000345332 9. Westwood DA, Eglinton TW, Frizelle FA (2011) Routine colonoscopy following acute uncomplicated diverticulitis. Br J Surg 98:1630–1634. doi:10.1002/bjs.7602
123
Surg Endosc (2014) 28:961–966 10. Sakhnini E, Lahat A, Melzer E et al (2004) Early colonoscopy in patients with acute diverticulitis: results of a prospective pilot study. Endoscopy 36:504–507. doi:10.1055/s-2004-814398 11. Lau KC, Spilsbury K, Farooque Y et al (2011) Is colonoscopy still mandatory after a CT diagnosis of left-sided diverticulitis: can colorectal cancer be confidently excluded? Dis Colon Rectum 54:1265–1270. doi:10.1097/DCR.0b013e31822899a2 12. Sai VF, Velayos F, Neuhaus J, Westphalen AC (2012) Colonoscopy after CT diagnosis of diverticulitis to exclude colon cancer: a systematic literature review. Radiology 263:383–390. doi:10.1148/radiol.12111869 13. Meurs-Szojda MM, Droste JSTS, Kuik DJ et al (2008) Diverticulosis and diverticulitis form no risk for polyps and colorectal neoplasia in 4,241 colonoscopies. Int J Colorectal Dis 23:979–984. doi:10.1007/s00384-008-0510-4 14. Lam TJ, Meurs-Szojda MM, Gundlach L et al (2009) There is no increased risk for colorectal cancer and adenomas in patients with diverticulitis: a retrospective longitudinal study. Colorectal Dis 12:1122–1126. doi:10.1111/j.1463-1318.2009.01992.x 15. Lahat A, Yanai H, Menachem Y et al (2007) The feasibility and risk of early colonoscopy in acute diverticulitis: a prospective controlled study. Endoscopy 39:521–524. doi:10.1055/s-2007-966399
