- Email: [email protected]
Actual endoscopic versus predicted surgical mortality for treatment of advanced mucosal neoplasia of the colon
ORIGINAL ARTICLE
Actual endoscopic versus predicted surgical mortality for treatment of advanced mucosal neoplasia of the colon Golo Ahlenstiel, MBBS, MD,1,7 Luke F. Hourigan, MBBS, FRACP,2,3 Gregor Brown, MBBS, FRACP, PhD,4,5 Simon Zanati, MBBS, FRACP,4,6 Stephen J. Williams, MBBS, FRACP, MD,1 Rajvinder Singh, MBBS, FRACP, MRCP, FRCP,6,7 Alan Moss, MBBS, FRACP, MD,1,6 Rebecca Sonson, BNurs,1 Michael J. Bourke, MBBS, FRACP,1,7 The Australian Colonic Endoscopic Mucosal Resection (ACE) Study Group Sydney, New South Wales; Brisbane, Queensland; Melbourne, Victoria; Adelaide, South Australia, Australia
Background: EMR of advanced mucosal neoplasia (AMN) (ie, sessile or laterally spreading lesions of R20 mm) of the colon has become an increasingly popular alternative to surgical resection. However, data regarding safety and mortality of EMR in comparison to surgery are limited. Objective: To compare actual endoscopic with predicted surgical mortality. Design: Prospective, observational, multicenter cohort study. Setting: Academic, high-volume, tertiary-care referral center. Patients: Consecutive patients referred for EMR. Intervention: EMR. Main Outcome Measurements: To predict hypothetical surgical mortality, the Association of Coloproctology of Great Britain and Ireland score, composed of physiological and surgical components, was calculated for each patient. Predicted surgical mortality was then compared with actual outcomes of EMR. The results were validated by an unselected subcohort by using the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity. Results: Among 1050 patients with AMN treated by EMR, including patients with a predicted mortality rate of greater than 5% (13.8% of cohort), no deaths occurred within 30 days after the procedure. The predicted surgical mortality rate was 3.3% with the Association of Coloproctology of Great Britain and Ireland score (P! .0001). This suggests a significant advantage of EMR over surgery. The results were validated by using the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity in 390 patients predicting a surgical mortality rate of 3.2% (P Z .0003). Limitations: Nonrandomized study. Conclusion: In this large multicenter study of EMR for colonic AMN, the predicted surgical mortality rate was significantly higher than the actual endoscopic mortality rate. Given that endoscopic therapy is less morbid and less expensive than surgery and can be performed as an outpatient treatment, it should be considered as the first line of treatment for most patients with these lesions. (Gastrointest Endosc 2014;-:1-9.)
Abbreviations: ACPGBI, Association of Coloproctology of Great Britain and Ireland; AMN, advanced mucosal neoplasia; ASA, American Society of Anesthesiologists; CRC, colorectal cancer; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; Hgb, hemoglobin; NNT, number needed to treat. DISCLOSURE: All authors disclosed no financial relationships relevant to this article. Copyright ª 2014 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2014.04.015
www.giejournal.org
Received December 22, 2013. Accepted April 2, 2014. Current affiliations: Department of Gastroenterology and Hepatology, Westmead Hospital, Sydney, New South Wales (1), Department of Gastroenterology and Hepatology, Princess Alexandra Hospital, Brisbane, Queensland (2), Department of Gastroenterology and Hepatology, Greenslopes Private Hospital, Brisbane, Queensland (3), Department of Gastroenterology and Hepatology, The Alfred Melbourne, Victoria (4), Department of Gastroenterology and Hepatology, Epworth Hospital, (footnotes continued on last page of article)
Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 1
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
Colorectal cancer (CRC) is the second most common cancer and cause of cancer death in Australia and the United States.1,2 In the majority of cases, it develops by the well-established adenoma-carcinoma sequence, allowing a therapeutic window to intervene. More than 85% of colonic polyps are small (!10 mm) and are easily treated by endoscopic polypectomy. Modeling against historical controls has clearly shown that colonoscopic polypectomy substantially reduces the incidence and mortality of CRC.3,4 With widespread CRC screening, increasing numbers of advanced mucosal neoplasias (AMNs) (ie, large [R20 mm] sessile polyps or laterally spreading tumors), are detected. These are premalignant and have an even stronger association with CRC than conventional polyps and therefore require therapy.5 Endoscopic resection of AMNs is technically challenging, time-consuming, and potentially hazardous. Thus, traditionally, surgery has been the standard of care and remains the most common treatment today.6-8 For example, in a recent report by the French CRC screening program, 9.3% of all individuals with a diagnosis of benign colonic adenoma underwent surgical resection.8 Colorectal surgery, however, is associated with significant peri- and postsurgical morbidity and mortality, particularly in these often elderly patients with multiple comorbidities.9,10 Over the past decade, EMR via ambulatory colonoscopy has become an increasingly used and viable therapeutic option (Figs. 1 and 2).11-13 Prospective studies have clearly shown the efficacy and safety of EMR for AMNs in an outpatient setting.12,14,15 Consequently, EMR has become established as an effective and safe therapy. Thus, it may not be ethically appropriate or logistically possible to perform a randomized, controlled trial against surgery. The best evidence in this context may therefore only come from modeling outcomes.3,4 The Association of Coloproctology of Great Britain and Ireland (ACPGBI) score and the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity (CR-POSSUM) were both developed and validated as dedicated user-friendly scoring systems to predict postoperative mortality in CRC (ACPGBI) or colorectal surgery (CR-POSSUM).16,17 Both combine measures of physiological health and operative severity with cancer stage to predict postoperative mortality.16,17 Numerous studies have confirmed the reliability of both scoring systems to predict 30-day mortality.17-23 The aim of the current study was to examine the safety of EMR versus surgery for AMNs by comparing actual endoscopic mortality with predicted surgical mortality in a large, prospective, multicenter cohort.
METHODS Study design and patient selection This study was part of a prospective, observational study of all patients referred for EMR of colonic AMNs 20 mm or 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
Take-home Message EMR for advanced mucosal neoplasia of the colon is safe and likely superior to surgery in terms of mortality, not only in healthy patients, but also in elderly patients with significant cardiac or renal disease. These results provide strong evidence that EMR, and not surgery, should be considered as the first line of treatment in the management of large colonic lesions.
larger at 7 Australian academic endoscopy units.12 Ethics approval was obtained by institutional review board approval for each participating center, and all patients gave written informed consent. Consecutive patients were enrolled from July 2008 to April 2012. A detailed description of the study protocol, including adverse events, was previously published.12 In brief, all patients were referred to a tertiary hospital for EMR after a previous colonoscopy by a nationally accredited consultant endoscopist had identified an advanced but ostensibly noninvasive lesion.11 Patients were reviewed by the investigators immediately before the procedure, and clinical follow-up was obtained at 14 days and subsequent intervals of 4 and 12 months. Endoscopic follow-up was obtained at intervals of 4 and 12 months. All authors had access to the study data and reviewed and approved the final manuscript. The EMR procedure was performed as previously described,11,12,24 unless there were features strongly suggestive of submucosal invasion, in which case, biopsy specimens were taken and surgical review was recommended. Patients with histological features of invasive malignancy in the EMR specimen were also referred for surgical review, and if they were unwilling or medically too unwell to have surgery, surveillance was offered.
Association of Coloproctology of Great Britain and Ireland score The ACPGBI consists of 2 physiological and 3 operative variables (Table 1). Mortality is then calculated based on R) the following formula16: LogR/(1 Z 4.859 þ total e score.
Data collection for ACPGBI Age and American Society of Anesthesiologists (ASA) grade were recorded prospectively per study protocol. The equivalent surgery was considered successful resection, as this would be the likely result of colorectal surgery of AMNs. The cancer stage was considered Dukes stage A and operative urgency elective. ASA grade was missing in 11 of 1061 patients, leaving 1050 patients with complete data for analysis. www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
Figure 1. EMR of large colonic mucosal neoplasia. Example of large (50 to 60 mm) hemicircumferential Paris Classification 0-IIaþIs laterally spreading tumor from the sigmoid colon before (A), during (B), and after (C) wide-field EMR.
Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity The CR-POSSUM includes 6 physiological and 5 operative variables (Table 2). Mortality is then calculated based on the following formula17: LogR/(1 R) Z 9.167 þ (0.338 physiological score) þ (0.308 operative score).
Data collection for CR-POSSUM To calculate the CR-POSSUM, the pulse rate and systolic blood pressure were assumed to be normal for all patients. If treated surgically, all lesions would have required abdominal surgery and thus classified as at least major in keeping with established criteria.17 Peritoneal soiling and Dukes stage were given the lowest possible score as nil/minor and A/B or less, respectively. All procedures were performed electively. Standard data collection in this multicenter study did not include hemoglobin (Hgb), urea, and the degree of cardiac failure specifically (although major cardiac comorbidities were recorded) and were therefore collected retrospectively for this prospective cohort, where this information was available. Blood test results for Hgb and urea were included if obtained within 3 months before the procedure. If patients were noted to have major cardiac comorbidities, medical records were reviewed to ascertain whether cardiac failure was present and its severity. To minimize selection bias, a second analysis was performed for the total cohort of 1061 patients recruited in www.giejournal.org
this time frame (referred to as the total cohort). In all cases without a urea value, urea was considered to be less than 10 mmol/L, and in all patients missing an Hgb value, Hgb was considered to be normal at 13.0 to 16.0 g/dL. Similarly, all patients with missing data regarding their cardiac status were considered to have no or only mild cardiac failure and a score of nil was assigned for this aspect.
Statistical analysis Predicted surgical mortality was calculated for each patient and then for the whole group relative to the number of patients per risk scores. The c2 or Fisher exact test were used where appropriate by using SPSS version 20.0 (IBM, Chicago, Ill). Two-sided P values !.05 were considered significant. All authors had access to the study data and reviewed and approved the final manuscript.
RESULTS This cohort of 1061 patients had a total of 1129 lesions with a mean size of 35.6 mm, with 54% in the right side of the colon. The mean age was 68 years (range 29-95 years). EMR was technically successful in 997 patients (91.5%). In 37 of 1061 patients (3.5%) EMR was not performed because of malignant features of the lesion (n Z 19, 1.8%) or technical difficulties (n Z 18, 1.7%). Seventy patients (6.6%) with technically successful EMR required surgical review subsequently for invasive disease demonstrated histologically in the EMR specimen. There were Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 3
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
Figure 2. Example of a large 50-60mm, hemi-circumferential Paris 0-IIa + Is laterally spreading tumor from the rectum. A, before, B, during and C, after wide-field endoscopic mucosal resection.
TABLE 1. Variables for ACPGBI score and associated points system17 Age, y
Score
Cancer
Score
Cancer stage
Score
Operative urgency
Score
Resected !65
0
ASA I
0
Dukes A
0
Elective
0
65-74
0.7
ASA II
0.8
Dukes B
0
Urgent
0.8
75-84
1.1
ASA III
1.6
Dukes C
0.2
Emergency
1.1
85-94
1.3
ASA IV-V
2.5
Dukes D
0.6
R95
2.6 Not resected ASA I
1.7
ASA II
1.8
ASA III
2.1
ASA IV-V
2.4
ASA, American Society of Anesthesiologists.
no deaths among the 1061 patients at 30 days after colonoscopy and/or EMR procedure.
ACPGBI score Complete data to calculate the ACPGBI score was prospectively collected and available for 1050 of 1061 patients (99.0%) undergoing colonoscopy with the intention of 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
EMR. The predominant ASA grades were I (healthy, 40.7%) and II (mild systemic disease not limiting activity, 46.3%) (Table 3). Mortality actual and predicted by ACPGBI score are listed in Table 4. Predicted surgical mortality was calculated to be 3.3%. Actual endoscopic mortality was 0.0% (P ! .0001) (Table 5). The number needed to treat (NNT) to avoid 1 death was 30. When only patients with www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
TABLE 2. Variables for CR-POSSUM and associated points system18 Score 1
2
3
4
8
!60
d
61-70
71-80
O81
None/mild
Moderate
Severe
d
d
Systolic BP, mm Hg
100-170
O170 or 90-99
!90
d
d
Pulse, beats/min
40-100
101-120
O120 or !40
d
d
Hemoglobin, g/dL
13-16
10-12.9 or 16.1-18
!10 or O18
d
d
Urea, mmol/L
%10
10.1-15.0
O15
d
d
Minor
d
Intermediate
Major
Complex major
Peritoneal soiling
None/serous fluid
Local pus
Free pus or feces
d
d
Operative urgency
Elective
d
Urgent
d
Emergency
None/Dukes A-B
Dukes C
Dukes
d
d
Physiological Age, y Cardiac failure
Operative score Operative severity
Cancer staging
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; BP, blood pressure.
successful EMR were included (n Z 917), the predicted surgical mortality and statistical significance remained the same (Table 5). When relatively well patients (ASA grade IþII) were analyzed separately, there was a significant difference in mortality in favor of EMR (n Z 911, 3.6% vs 0%; P ! .0001). The same was true for patients with an ASA grade of III or higher (n Z 137, 8.7% vs 0%; P Z .0004).
avoided by using EMR in an intention-to-treat analysis (Table 5). When only successful EMR procedures were considered, these numbers were 11 (CR-POSSUM, P Z .0008) or 22 (total cohort, P ! .0001). This translates to an NNT with EMR instead of surgery of 30 or 42, respectively.
Comorbidity Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity To validate our results, we used the CR-POSSUM, which is dependent on variables different from those for the ACPGBI score. Complete data (including Hgb, urea, and cardiac failure status) to calculate the CR-POSSUM was available for 390 patients (36.8% of total cohort; CRPOSSUM cohort). As data required for CR-POSSUM were collected retrospectively, it was incomplete for the remaining 671 patients (63.2%) (ie, missing Hgb for 658 patients [62%], urea for 665 patients [62.7%], and cardiac failure status for 671 patients [63.2%]). To minimize selection bias, all analyses were also performed for the whole cohort by assuming the lowest possible scores and thus the lowest mortality (see Methods section). The baseline statistics for the CR-POSSUM cohort are provided in Table 3. Surgical mortality according to the CR-POSSUM was calculated at 3.2% in the CR-POSSUM cohort and 2.4% across the total cohort (Table 5). As actual endoscopic mortality was 0% (P Z .0003 for the CR-POSSUM cohort or P ! .0001 for total cohort), an estimated 13 (CR-POSSUM cohort) or 25 potential perioperative deaths were www.giejournal.org
Importantly, the study included a significant number of high-risk patients: 13.8% per the ACPGBI score and 17.4% per the CR-POSSUM had a predicted mortality of greater than 5% (Tables 4 and 6), which would translate into an NNT less than 20. This held true for patients with impaired renal function (urea O10 mmol/L) and thus likely numerous other significant medical problems not covered by ACPGBI or CR-POSSUM (n Z 41) (Table 7).
DISCUSSION The relative distribution of colonic AMN to either surgical or endoscopic management is unknown; however, it seems likely that many patients who could be treated efficiently, safely, and cost-effectively by EMR are still being offered only surgery. For example in a recent retrospective review, 27 of 38 advanced colonic lesions without biopsyproven cancer referred for surgical resection could be safely removed by endoscopic resection, avoiding surgery completely.25 If it were not for an enlightened surgeon, all of these patients would indeed have had surgery. Further evidence to support the role of endoscopic Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 5
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
TABLE 3. Baseline characteristics Characteristics
TABLE 4. Mortality stratified by ACPGBI score Overall (N [ 1061)
No. (%)
Actual endoscopic mortality
0.0
227 (21.6)
0 (0.0)
0.8
ACPGBI subcohort (N [ 1050)
0.7
147 (14.0)
0 (0.0)
1.5
ASA grade, no. (%)
0.8
108 (10.3)
0 (0.0)
1.7
Sex, male:female ratio) Age, y (range)
1.13:1 68 (27-95)
ACPGBI summary score
Predicted surgical mortality, %
I
427 (40.7)
1.1
51 (4.9)
0 (0.0)
2.3
II
486 (46.3)
1.3
3 (0.3)
0 (0.0)
2.8
III
132 (12.6)
1.5
181 (17.2)
0 (0.0)
3.4
IV
5 (0.5)
1.6
24 (2.3)
0 (0.0)
3.7
CR-POSSUM subcohort (N [ 390)
1.9
164 (15.6)
0 (0.0)
4.9
Hgb (range), g/L
12.2 (6.0-18.0)
2.1
31 (3.0)
0 (0.0)
6.0
Urea (range), mmol/L
6.4 (1.0-38.0)
2.3
39 (3.7)
0 (0.0)
7.2
Cardiac failure, no. (%)
2.5
1 (0.1)
0 (0.0)
8.6
None/mild
354 (90.8)
2.7
59 (5.6)
0 (0.0)
10.3
Moderate
27 (6.9)
2.9
10 (1.0)
0 (0.0)
12.4
Severe
9 (2.3)
3.2
1 (0.1)
0 (0.0)
16.0
3.4
1 (0.1)
0 (0.0)
18.9
3.6
3 (0.3)
0 (0.0)
22.1
ACPGBI, Association of Coloproctology of Great Britain and Ireland; ASA, American Society of Anesthesiologists; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; Hgb, hemoglobin.
ACPGBI, Association of Coloproctology of Great Britain and Ireland.
resection is still required, and comparative safety data are still largely lacking. In this large, prospective, multicenter cohort of patients with colonic AMNs treated by EMR, endoscopic mortality was 0 and significantly less than surgical mortality as predicted by 2 well-validated scoring systems. Previous studies on colorectal surgery outcomes for patients with CRC have shown that the ACPGBI score18,21,22,26 and CR-POSSUM accurately predict surgical mortality.19,21,22,27-30 Furthermore, where data were incomplete, we adopted a conservative model that favored the surgical approach by assuming normal scores for these patients, thus likely artificially reducing the predicted surgical mortality. Even so, there was a significant advantage of actual endoscopic mortality over predicted surgical mortality. We used 2 different risk scores to predict surgical mortality in our patient cohort: 1 that was designed for patients with CRC (ACPGBI score) and 1 that was designed for colorectal surgery including benign pathology (CR-POSSUM). It is important to note that both scoring systems predicted a similar mortality of 3.3% and 3.2%, respectively. As both scoring systems use quite different physiological variables apart from age, the similar predicted mortality is likely to be valid. Importantly, the actual endoscopic mortality rate was significantly lower than the predicted surgical mortality rate, even in relatively healthy patients (ASA 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
grades I and II). Even if we assume that predicted surgical mortality was overestimated, 2 large prospective case series with 750 consecutive patients undergoing laparoscopic colonic or rectal surgery31 and 15,427 patients undergoing open colonic segmental resection32 found mortality rates of 2.2% and 4.4%, respectively, both still significantly higher than the endoscopic mortality rate in our cohort.31,32 CR-POSSUM scoring was used on an unselected subcohort, which confirmed results from ACPGBI scoring. To minimize selection bias through better recall, ie, better availability of blood tests and cardiac history on more debilitated patients, we assumed lowest possible CR-POSSUM and thus lowest mortality rate for patients with missing data in the analysis of the total cohort. Thus, overall mortality per the CR-POSSUM in the whole study cohort was rather underestimated than overestimated with nevertheless a highly significant difference between actual endoscopic and predicted surgical mortality. This study was unique in several aspects. It is by far the largest, prospectively collected patient cohort resulting in good statistical power, especially as all referred patients were included without exceptions. Our study was designed as a multicenter study in 7 separate, academic tertiary centers in Australia, collecting prospectively a large group of www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
TABLE 5. Actual endoscopic and predicted surgical mortality No.
AEM, no. (%)
PSM, no. (%)
OR (CI)
RR (CI)
P value
NNT
All patients
1050
0 (0)
35 (3.3)
0 (0-NaN)
0.4915 (0.4704-0.5136)
!.0001
30
Successful EMR
917
0 (0)
30 (3.3)
0 (0-NaN)
0.4917 (0.4691-0.5153)
!.0001
30
All patients
390
0 (0)
13 (3.2)
0 (0-NaN)
0.4915 (0.4574-0.5282)
.0003
30
Successful EMR
336
0 (0)
11 (3.3)
0 (0-NaN)
0.4917 (0.4550-0.5313)
.0009
31
All patients
1061
0 (0)
25 (2.4)
0 (0-NaN)
0.494 (0.4731-0.5159)
!.0001
42
Successful EMR
921
0 (0)
22 (2.4)
0 (0-NaN)
0.494 (0.4715-0.5175)
!.0001
42
Group ACPGBI (complete data)
CR-POSSUM (complete data)
CR-POSSUM (total cohort)
AEM, Actual endoscopic mortality; PSM, predicted surgical mortality; OR, odds ratio; CI, confidence interval; RR, risk ratio; NNT, number needed to treat; ACPGBI, Association of Coloproctology of Great Britain and Ireland; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; NaN, not a number.
TABLE 6. Mortality stratified by CR-POSSUM
TABLE 7. Predicted surgical mortality by CR-POSSUM stratified by urea
No. (%)
Actual endoscopic mortality
Predicted surgical mortality, %
6
46 (11.8)
0 (0.0)
0.7
7
35 (9.0)
0 (0.0)
1.0
8
69 (17.7)
0 (0.0)
1.3
9
93 (23.8)
0 (0.0)
1.9
10
57 (14.6)
0 (0.0)
2.6
11
22 (5.6)
0 (0.0)
3.6
12
3 (0.8)
0 (0.0)
5.0
13
19 (4.9)
0 (0.0)
6.9
14
27 (6.9)
0 (0.0)
9.3
15
8 (2.1)
0 (0.0)
12.6
16
7 (1.8)
0 (0.0)
16.8
17
2 (0.5)
0 (0.0)
22.0
18
2 (0.5)
0 (0.0)
28.4
Physiological score
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity.
patients undergoing EMR for AMNs within a reasonably short time (34 months). This ensured continuity in terms of equipment standards, endoscopic procedures, and protocol, as well as consultant endoscopists performing the procedures. Also data collection was comprehensive and follow-up was complete. The best evidence to confirm a 30-day mortality advantage of EMR against surgery would undoubtedly be derived www.giejournal.org
Predicted surgical mortality
Urea, no. (%) %10
10.1-15.0
O15.0
0.7
46 (13.2)
0 (0.0)
0 (0.0)
1.0
35 (10.0)
0 (0.0)
0 (0.0)
1.3
68 (19.5)
1 (3.4)
0 (0.0)
1.9
90 (25.8)
1 (3.4)
2 (16.7)
2.6
51 (14.6)
5 (17.2)
1 (8.3)
3.6
9 (2.6)
13 (44.8)
0 (0.0)
5.0
0 (0.0)
0 (0.0)
3 (25.0)
6.9
18 (5.2)
1 (3.4)
0 (0.0)
9.3
27 (7.7)
0 (0.0)
0 (0.0)
12.6
4 (1.1)
4 (13.8)
0 (0.0)
16.8
1 (0.3)
4 (13.8)
2 (16.7)
22.0
0 (0.0)
0 (0.0)
2 (16.7)
28.4
0 (0.0)
0 (0.0)
2 (16.7)
Total
349
29
12
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity.
from a multicenter, randomized, controlled trial. However, such a study would seem potentially unethical and logistically very challenging to perform. EMR has proven itself to be a safe and effective therapy for AMN in patients with a predicted 30-day mortality from elective surgery of more Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 7
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
than 5% or more than 10% (comprising 13.8% and 7% of our cohort, respectively, by ACPGBI) could not be ethically assigned to the surgery arm. Given the large number of prospectively recruited patients, comorbidities, and age in the study, this cohort is likely to be representative of patients needing treatment for colonic AMN. Nevertheless, it is difficult to rule out a selection bias with our patients negatively preselected toward comorbidity, ie, their comorbidities may be a reason why they were referred for EMR rather than surgery. However, 87% of the group were ASA I or II, and even in these relatively healthy patients, the 30-day mortality was significantly lower for EMR than predicted for surgery, suggesting that EMR is superior even in relatively well patients. Beyond perioperative mortality risk, both the endoscopic and the surgical approaches have their own benefit and risk profile. Surgery is usually definitive, but carries the potential of major morbidity in 9%33 and anastomotic leak in 6% to 7% with extended hospital stay.34 The main adverse event with EMR is significant postprocedural bleeding necessitating hospital admission, which occurs in 6% to 7% overall.35 Perforation is infrequent at 1% to 2% and can now be recognized and managed effectively endoscopically in the majority of patients without major sequelae.24 The major limitation of EMR is the need for scheduled colonoscopic surveillance at 4 to 6 months after EMR. There is recurrence in approximately 15% to 20%, but it is usually minor, unifocal, and easily treated. Finally, with increasing life expectancy, improving management of comorbidities worldwide, and increased uptake of CRC screening, the necessity to remove premalignant lesions in elderly patients with substantial comorbidities is likely to grow. A good example is the patient with renal disease in whom cancer is the second most common cause of death.36 Our study included a significant number of patients with renal disease (Table 7) and no mortality from EMR. Currently, the limitations of an EMR strategy for the treatment of AMN is the need for an experienced endoscopist, a dedicated team, suitable endoscopic equipment, and surgical support, a combination usually only available in tertiary centers and therefore currently not yet readily accessible everywhere. Nevertheless, our study underlines the need to establish more centers capable of performing safe and effective EMR, considering that demand is increasing with an aging population with increasing comorbidities. In summary, the data from this large multicenter cohort suggest that EMR is not only noninferior, but is safer than traditionally accepted surgical resection for the treatment of colonic AMN. Given that endoscopic therapy is less morbid and expensive than surgery and can be performed as an outpatient, endoscopic resection should be considered the first-line treatment for colonic AMNs. 8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
REFERENCES 1. AIHW, CA (Cancer Australia) & AAACR (Australasian Association of Cancer Registries), 2008 R. Cancer survival and prevalence in Australia: cancers diagnosed from 1982 to 2004. Cancer series no. 42. Cat. No. CAN 38. Canberra (Australia): AIHW. 2. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 1975;116:544-73. 3. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977-81. 4. Zauber AG, Winawer SJ, O'Brien MJ, et al. Colonoscopic polypectomy and long-term prevention of colorectal-cancer deaths. N Engl J Med 2012;366:687-96. 5. Soetikno RM, Kaltenbach T, Rouse RV, et al. Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults. JAMA 2008;299:1027-35. 6. Voloyiannis T, Snyder MJ, Bailey RR, et al. Management of the difficult colon polyp referred for resection: resect or rescope? Dis Colon Rectum 2008;51:292-5. 7. Onken JE, Friedman JY, Subramanian S, et al. Treatment patterns and costs associated with sessile colorectal polyps. Am J Gastroenterol 2002;97:2896-901. 8. Manfredi S, Piette C, Durand G, et al. Colonoscopy results of a French regional FOBT-based colorectal cancer screening program with high compliance. Endoscopy 2008;40:422-7. 9. McNicol L, Story DA, Leslie K, et al. Postoperative complications and mortality in older patients having non-cardiac surgery at three Melbourne teaching hospitals. Med J Aust 2007;186:447-52. 10. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128-37. 11. Swan MP, Bourke MJ, Alexander S, et al. Large refractory colonic polyps: is it time to change our practice? A prospective study of the clinical and economic impact of a tertiary referral colonic mucosal resection and polypectomy service (with videos). Gastrointest Endosc 2009;70:1128-36. 12. Moss A, Bourke MJ, Williams SJ, et al. Endoscopic mucosal resection outcomes and prediction of submucosal cancer from advanced colonic mucosal neoplasia. Gastroenterology 2011;140:1909-18. 13. Uraoka T, Saito Y, Matsuda T, et al. Endoscopic indications for endoscopic mucosal resection of laterally spreading tumours in the colorectum. Gut 2006;55:1592-7. 14. Conio M, Repici A, Demarquay JF, et al. EMR of large sessile colorectal polyps. Gastrointest Endosc 2004;60:234-41. 15. Buchner AM, Guarner-Argente C, Ginsberg GG. Outcomes of EMR of defiant colorectal lesions directed to an endoscopy referral center. Gastrointest Endosc 2012;76:255-63. 16. Tekkis PP, Poloniecki JD, Thompson MR, et al. Operative mortality in colorectal cancer: prospective national study. BMJ 2003;327:1196-201. 17. Tekkis PP, Prytherch DR, Kocher HM, et al. Development of a dedicated risk-adjustment scoring system for colorectal surgery (colorectal POSSUM). Br J Surg 2004;91:1174-82. 18. Al-Homoud S, Purkayastha S, Aziz O, et al. Evaluating operative risk in colorectal cancer surgery: ASA and POSSUM-based predictive models. Surg Oncol 2004;13:83-92. 19. Horzic M, Kopljar M, Cupurdija K, et al. Comparison of P-POSSUM and Cr-POSSUM scores in patients undergoing colorectal cancer resection. Arch Surg 2007;142:1043-8. 20. Vather R, Zargar-Shoshtari K, Adegbola S, et al. Comparison of the possum, P-POSSUM and Cr-POSSUM scoring systems as predictors of postoperative mortality in patients undergoing major colorectal surgery. ANZ J Surg 2006;76:812-6. 21. Yan J, Wang YX, Li ZP. Predictive value of the POSSUM, p-POSSUM, crPOSSUM, APACHE II and ACPGBI scoring systems in colorectal cancer resection. J Int Med Res 2011;39:1464-73.
www.giejournal.org
Ahlenstiel et al 22. Ferjani AM, Griffin D, Stallard N, et al. A newly devised scoring system for prediction of mortality in patients with colorectal cancer: a prospective study. Lancet Oncol 2007;8:317-22. 23. Senagore AJ, Warmuth AJ, Delaney CP, et al. POSSUM, p-POSSUM, and Cr-POSSUM: implementation issues in a United States health care system for prediction of outcome for colon cancer resection. Dis Colon Rectum 2004;47:1435-41. 24. Holt BA, Bourke MJ. Wide field endoscopic resection for advanced colonic mucosal neoplasia: current status and future directions. Clin Gastroenterol Hepatol 2012;10:969-79. 25. Friedland S, Banerjee S, Kochar R, et al. Outcomes of repeat colonoscopy in patients with polyps referred for surgery without biopsyproven cancer. Gastrointest Endosc 2013;79:101-7. 26. Teeuwen PH, Bremers AJ, Groenewoud JM, et al. Predictive value of POSSUM and ACPGBI scoring in mortality and morbidity of colorectal resection: a case-control study. J Gastrointest Surg 2011;15: 294-303. 27. Bromage SJ, Cunliffe WJ. Validation of the CR-POSSUM risk-adjusted scoring system for major colorectal cancer surgery in a single center. Dis Colon Rectum 2007;50:192-6. 28. Leung E, Ferjani AM, Stellard N, et al. Predicting post-operative mortality in patients undergoing colorectal surgery using P-POSSUM and CRPOSSUM scores: a prospective study. Int J Colorectal Dis 2009;24: 1459-64. 29. Leung E, Ferjani AM, Kitchen A, et al. Risk-adjusted scoring systems can predict surgeons’ performance in colorectal surgery. Surgeon 2011;9: 3-7. 30. Ptok H, Marusch F, Schmidt U, et al. Risk adjustment as basis for rational benchmarking: the example of colon carcinoma. World J Surg 2011;35:196-205.
www.giejournal.org
Endoscopic mortality for EMR of colonic AMN 31. Schlachta CM, Mamazza J, Gregoire R, et al. Could laparoscopic colon and rectal surgery become the standard of care? A review and experience with 750 procedures. Can J Surg 2003;46:432-40. 32. Prystowsky JB, Bordage G, Feinglass JM. Patient outcomes for segmental colon resection according to surgeon’s training, certification, and experience. Surgery 2002;132:663-70. 33. Cirocchi R, Farinella E, Trastulli S, et al. Elective sigmoid colectomy for diverticular disease. Laparoscopic vs open surgery: a systematic review. Colorectal Dis 2012;14:671-83. 34. Krarup PM, Jorgensen LN, Andreasen AH, et al. A nationwide study on anastomotic leakage after colonic cancer surgery. Colorectal Dis 2012;14:e661-7. 35. Burgess NG, Metz AJ, Williams SJ, et al. Risk factors for intraprocedural and clinically significant delayed Bleeding after wide-field endoscopic mucosal resection of large colonic lesions. Clin Gastroenterol Hepatol 2014;12:651-61.e3. 36. Australian and New Zealand Dialysis and Transplant Registry, The 30th Annual Report, 2007 Available at http://www.anzdata.org.au/v1/ annual_reports_download.html. Melbourne, Victoria (5), Department of Gastroenterology and Hepatology, Western Hospital, Melbourne, Victoria (6), Department of Gastroenterology and Hepatology, Lyell McEwin Hospital, Adelaide, South Australia (7), Australia. Michael J. Bourke, MBBS, FRACP, Department of Gastroenterology and Hepatology, Westmead Hospital, Hawkesbury Road, Westmead, NSW 2145, Australia. If you would like to chat with an author of this article, you may contact Dr Ahlenstiel at [email protected].
Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 9
Actual endoscopic versus predicted surgical mortality for treatment of advanced mucosal neoplasia of the colon Golo Ahlenstiel, MBBS, MD,1,7 Luke F. Hourigan, MBBS, FRACP,2,3 Gregor Brown, MBBS, FRACP, PhD,4,5 Simon Zanati, MBBS, FRACP,4,6 Stephen J. Williams, MBBS, FRACP, MD,1 Rajvinder Singh, MBBS, FRACP, MRCP, FRCP,6,7 Alan Moss, MBBS, FRACP, MD,1,6 Rebecca Sonson, BNurs,1 Michael J. Bourke, MBBS, FRACP,1,7 The Australian Colonic Endoscopic Mucosal Resection (ACE) Study Group Sydney, New South Wales; Brisbane, Queensland; Melbourne, Victoria; Adelaide, South Australia, Australia
Background: EMR of advanced mucosal neoplasia (AMN) (ie, sessile or laterally spreading lesions of R20 mm) of the colon has become an increasingly popular alternative to surgical resection. However, data regarding safety and mortality of EMR in comparison to surgery are limited. Objective: To compare actual endoscopic with predicted surgical mortality. Design: Prospective, observational, multicenter cohort study. Setting: Academic, high-volume, tertiary-care referral center. Patients: Consecutive patients referred for EMR. Intervention: EMR. Main Outcome Measurements: To predict hypothetical surgical mortality, the Association of Coloproctology of Great Britain and Ireland score, composed of physiological and surgical components, was calculated for each patient. Predicted surgical mortality was then compared with actual outcomes of EMR. The results were validated by an unselected subcohort by using the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity. Results: Among 1050 patients with AMN treated by EMR, including patients with a predicted mortality rate of greater than 5% (13.8% of cohort), no deaths occurred within 30 days after the procedure. The predicted surgical mortality rate was 3.3% with the Association of Coloproctology of Great Britain and Ireland score (P! .0001). This suggests a significant advantage of EMR over surgery. The results were validated by using the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity in 390 patients predicting a surgical mortality rate of 3.2% (P Z .0003). Limitations: Nonrandomized study. Conclusion: In this large multicenter study of EMR for colonic AMN, the predicted surgical mortality rate was significantly higher than the actual endoscopic mortality rate. Given that endoscopic therapy is less morbid and less expensive than surgery and can be performed as an outpatient treatment, it should be considered as the first line of treatment for most patients with these lesions. (Gastrointest Endosc 2014;-:1-9.)
Abbreviations: ACPGBI, Association of Coloproctology of Great Britain and Ireland; AMN, advanced mucosal neoplasia; ASA, American Society of Anesthesiologists; CRC, colorectal cancer; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; Hgb, hemoglobin; NNT, number needed to treat. DISCLOSURE: All authors disclosed no financial relationships relevant to this article. Copyright ª 2014 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2014.04.015
www.giejournal.org
Received December 22, 2013. Accepted April 2, 2014. Current affiliations: Department of Gastroenterology and Hepatology, Westmead Hospital, Sydney, New South Wales (1), Department of Gastroenterology and Hepatology, Princess Alexandra Hospital, Brisbane, Queensland (2), Department of Gastroenterology and Hepatology, Greenslopes Private Hospital, Brisbane, Queensland (3), Department of Gastroenterology and Hepatology, The Alfred Melbourne, Victoria (4), Department of Gastroenterology and Hepatology, Epworth Hospital, (footnotes continued on last page of article)
Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 1
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
Colorectal cancer (CRC) is the second most common cancer and cause of cancer death in Australia and the United States.1,2 In the majority of cases, it develops by the well-established adenoma-carcinoma sequence, allowing a therapeutic window to intervene. More than 85% of colonic polyps are small (!10 mm) and are easily treated by endoscopic polypectomy. Modeling against historical controls has clearly shown that colonoscopic polypectomy substantially reduces the incidence and mortality of CRC.3,4 With widespread CRC screening, increasing numbers of advanced mucosal neoplasias (AMNs) (ie, large [R20 mm] sessile polyps or laterally spreading tumors), are detected. These are premalignant and have an even stronger association with CRC than conventional polyps and therefore require therapy.5 Endoscopic resection of AMNs is technically challenging, time-consuming, and potentially hazardous. Thus, traditionally, surgery has been the standard of care and remains the most common treatment today.6-8 For example, in a recent report by the French CRC screening program, 9.3% of all individuals with a diagnosis of benign colonic adenoma underwent surgical resection.8 Colorectal surgery, however, is associated with significant peri- and postsurgical morbidity and mortality, particularly in these often elderly patients with multiple comorbidities.9,10 Over the past decade, EMR via ambulatory colonoscopy has become an increasingly used and viable therapeutic option (Figs. 1 and 2).11-13 Prospective studies have clearly shown the efficacy and safety of EMR for AMNs in an outpatient setting.12,14,15 Consequently, EMR has become established as an effective and safe therapy. Thus, it may not be ethically appropriate or logistically possible to perform a randomized, controlled trial against surgery. The best evidence in this context may therefore only come from modeling outcomes.3,4 The Association of Coloproctology of Great Britain and Ireland (ACPGBI) score and the Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity (CR-POSSUM) were both developed and validated as dedicated user-friendly scoring systems to predict postoperative mortality in CRC (ACPGBI) or colorectal surgery (CR-POSSUM).16,17 Both combine measures of physiological health and operative severity with cancer stage to predict postoperative mortality.16,17 Numerous studies have confirmed the reliability of both scoring systems to predict 30-day mortality.17-23 The aim of the current study was to examine the safety of EMR versus surgery for AMNs by comparing actual endoscopic mortality with predicted surgical mortality in a large, prospective, multicenter cohort.
METHODS Study design and patient selection This study was part of a prospective, observational study of all patients referred for EMR of colonic AMNs 20 mm or 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
Take-home Message EMR for advanced mucosal neoplasia of the colon is safe and likely superior to surgery in terms of mortality, not only in healthy patients, but also in elderly patients with significant cardiac or renal disease. These results provide strong evidence that EMR, and not surgery, should be considered as the first line of treatment in the management of large colonic lesions.
larger at 7 Australian academic endoscopy units.12 Ethics approval was obtained by institutional review board approval for each participating center, and all patients gave written informed consent. Consecutive patients were enrolled from July 2008 to April 2012. A detailed description of the study protocol, including adverse events, was previously published.12 In brief, all patients were referred to a tertiary hospital for EMR after a previous colonoscopy by a nationally accredited consultant endoscopist had identified an advanced but ostensibly noninvasive lesion.11 Patients were reviewed by the investigators immediately before the procedure, and clinical follow-up was obtained at 14 days and subsequent intervals of 4 and 12 months. Endoscopic follow-up was obtained at intervals of 4 and 12 months. All authors had access to the study data and reviewed and approved the final manuscript. The EMR procedure was performed as previously described,11,12,24 unless there were features strongly suggestive of submucosal invasion, in which case, biopsy specimens were taken and surgical review was recommended. Patients with histological features of invasive malignancy in the EMR specimen were also referred for surgical review, and if they were unwilling or medically too unwell to have surgery, surveillance was offered.
Association of Coloproctology of Great Britain and Ireland score The ACPGBI consists of 2 physiological and 3 operative variables (Table 1). Mortality is then calculated based on R) the following formula16: LogR/(1 Z 4.859 þ total e score.
Data collection for ACPGBI Age and American Society of Anesthesiologists (ASA) grade were recorded prospectively per study protocol. The equivalent surgery was considered successful resection, as this would be the likely result of colorectal surgery of AMNs. The cancer stage was considered Dukes stage A and operative urgency elective. ASA grade was missing in 11 of 1061 patients, leaving 1050 patients with complete data for analysis. www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
Figure 1. EMR of large colonic mucosal neoplasia. Example of large (50 to 60 mm) hemicircumferential Paris Classification 0-IIaþIs laterally spreading tumor from the sigmoid colon before (A), during (B), and after (C) wide-field EMR.
Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity The CR-POSSUM includes 6 physiological and 5 operative variables (Table 2). Mortality is then calculated based on the following formula17: LogR/(1 R) Z 9.167 þ (0.338 physiological score) þ (0.308 operative score).
Data collection for CR-POSSUM To calculate the CR-POSSUM, the pulse rate and systolic blood pressure were assumed to be normal for all patients. If treated surgically, all lesions would have required abdominal surgery and thus classified as at least major in keeping with established criteria.17 Peritoneal soiling and Dukes stage were given the lowest possible score as nil/minor and A/B or less, respectively. All procedures were performed electively. Standard data collection in this multicenter study did not include hemoglobin (Hgb), urea, and the degree of cardiac failure specifically (although major cardiac comorbidities were recorded) and were therefore collected retrospectively for this prospective cohort, where this information was available. Blood test results for Hgb and urea were included if obtained within 3 months before the procedure. If patients were noted to have major cardiac comorbidities, medical records were reviewed to ascertain whether cardiac failure was present and its severity. To minimize selection bias, a second analysis was performed for the total cohort of 1061 patients recruited in www.giejournal.org
this time frame (referred to as the total cohort). In all cases without a urea value, urea was considered to be less than 10 mmol/L, and in all patients missing an Hgb value, Hgb was considered to be normal at 13.0 to 16.0 g/dL. Similarly, all patients with missing data regarding their cardiac status were considered to have no or only mild cardiac failure and a score of nil was assigned for this aspect.
Statistical analysis Predicted surgical mortality was calculated for each patient and then for the whole group relative to the number of patients per risk scores. The c2 or Fisher exact test were used where appropriate by using SPSS version 20.0 (IBM, Chicago, Ill). Two-sided P values !.05 were considered significant. All authors had access to the study data and reviewed and approved the final manuscript.
RESULTS This cohort of 1061 patients had a total of 1129 lesions with a mean size of 35.6 mm, with 54% in the right side of the colon. The mean age was 68 years (range 29-95 years). EMR was technically successful in 997 patients (91.5%). In 37 of 1061 patients (3.5%) EMR was not performed because of malignant features of the lesion (n Z 19, 1.8%) or technical difficulties (n Z 18, 1.7%). Seventy patients (6.6%) with technically successful EMR required surgical review subsequently for invasive disease demonstrated histologically in the EMR specimen. There were Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 3
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
Figure 2. Example of a large 50-60mm, hemi-circumferential Paris 0-IIa + Is laterally spreading tumor from the rectum. A, before, B, during and C, after wide-field endoscopic mucosal resection.
TABLE 1. Variables for ACPGBI score and associated points system17 Age, y
Score
Cancer
Score
Cancer stage
Score
Operative urgency
Score
Resected !65
0
ASA I
0
Dukes A
0
Elective
0
65-74
0.7
ASA II
0.8
Dukes B
0
Urgent
0.8
75-84
1.1
ASA III
1.6
Dukes C
0.2
Emergency
1.1
85-94
1.3
ASA IV-V
2.5
Dukes D
0.6
R95
2.6 Not resected ASA I
1.7
ASA II
1.8
ASA III
2.1
ASA IV-V
2.4
ASA, American Society of Anesthesiologists.
no deaths among the 1061 patients at 30 days after colonoscopy and/or EMR procedure.
ACPGBI score Complete data to calculate the ACPGBI score was prospectively collected and available for 1050 of 1061 patients (99.0%) undergoing colonoscopy with the intention of 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
EMR. The predominant ASA grades were I (healthy, 40.7%) and II (mild systemic disease not limiting activity, 46.3%) (Table 3). Mortality actual and predicted by ACPGBI score are listed in Table 4. Predicted surgical mortality was calculated to be 3.3%. Actual endoscopic mortality was 0.0% (P ! .0001) (Table 5). The number needed to treat (NNT) to avoid 1 death was 30. When only patients with www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
TABLE 2. Variables for CR-POSSUM and associated points system18 Score 1
2
3
4
8
!60
d
61-70
71-80
O81
None/mild
Moderate
Severe
d
d
Systolic BP, mm Hg
100-170
O170 or 90-99
!90
d
d
Pulse, beats/min
40-100
101-120
O120 or !40
d
d
Hemoglobin, g/dL
13-16
10-12.9 or 16.1-18
!10 or O18
d
d
Urea, mmol/L
%10
10.1-15.0
O15
d
d
Minor
d
Intermediate
Major
Complex major
Peritoneal soiling
None/serous fluid
Local pus
Free pus or feces
d
d
Operative urgency
Elective
d
Urgent
d
Emergency
None/Dukes A-B
Dukes C
Dukes
d
d
Physiological Age, y Cardiac failure
Operative score Operative severity
Cancer staging
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; BP, blood pressure.
successful EMR were included (n Z 917), the predicted surgical mortality and statistical significance remained the same (Table 5). When relatively well patients (ASA grade IþII) were analyzed separately, there was a significant difference in mortality in favor of EMR (n Z 911, 3.6% vs 0%; P ! .0001). The same was true for patients with an ASA grade of III or higher (n Z 137, 8.7% vs 0%; P Z .0004).
avoided by using EMR in an intention-to-treat analysis (Table 5). When only successful EMR procedures were considered, these numbers were 11 (CR-POSSUM, P Z .0008) or 22 (total cohort, P ! .0001). This translates to an NNT with EMR instead of surgery of 30 or 42, respectively.
Comorbidity Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity To validate our results, we used the CR-POSSUM, which is dependent on variables different from those for the ACPGBI score. Complete data (including Hgb, urea, and cardiac failure status) to calculate the CR-POSSUM was available for 390 patients (36.8% of total cohort; CRPOSSUM cohort). As data required for CR-POSSUM were collected retrospectively, it was incomplete for the remaining 671 patients (63.2%) (ie, missing Hgb for 658 patients [62%], urea for 665 patients [62.7%], and cardiac failure status for 671 patients [63.2%]). To minimize selection bias, all analyses were also performed for the whole cohort by assuming the lowest possible scores and thus the lowest mortality (see Methods section). The baseline statistics for the CR-POSSUM cohort are provided in Table 3. Surgical mortality according to the CR-POSSUM was calculated at 3.2% in the CR-POSSUM cohort and 2.4% across the total cohort (Table 5). As actual endoscopic mortality was 0% (P Z .0003 for the CR-POSSUM cohort or P ! .0001 for total cohort), an estimated 13 (CR-POSSUM cohort) or 25 potential perioperative deaths were www.giejournal.org
Importantly, the study included a significant number of high-risk patients: 13.8% per the ACPGBI score and 17.4% per the CR-POSSUM had a predicted mortality of greater than 5% (Tables 4 and 6), which would translate into an NNT less than 20. This held true for patients with impaired renal function (urea O10 mmol/L) and thus likely numerous other significant medical problems not covered by ACPGBI or CR-POSSUM (n Z 41) (Table 7).
DISCUSSION The relative distribution of colonic AMN to either surgical or endoscopic management is unknown; however, it seems likely that many patients who could be treated efficiently, safely, and cost-effectively by EMR are still being offered only surgery. For example in a recent retrospective review, 27 of 38 advanced colonic lesions without biopsyproven cancer referred for surgical resection could be safely removed by endoscopic resection, avoiding surgery completely.25 If it were not for an enlightened surgeon, all of these patients would indeed have had surgery. Further evidence to support the role of endoscopic Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 5
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
TABLE 3. Baseline characteristics Characteristics
TABLE 4. Mortality stratified by ACPGBI score Overall (N [ 1061)
No. (%)
Actual endoscopic mortality
0.0
227 (21.6)
0 (0.0)
0.8
ACPGBI subcohort (N [ 1050)
0.7
147 (14.0)
0 (0.0)
1.5
ASA grade, no. (%)
0.8
108 (10.3)
0 (0.0)
1.7
Sex, male:female ratio) Age, y (range)
1.13:1 68 (27-95)
ACPGBI summary score
Predicted surgical mortality, %
I
427 (40.7)
1.1
51 (4.9)
0 (0.0)
2.3
II
486 (46.3)
1.3
3 (0.3)
0 (0.0)
2.8
III
132 (12.6)
1.5
181 (17.2)
0 (0.0)
3.4
IV
5 (0.5)
1.6
24 (2.3)
0 (0.0)
3.7
CR-POSSUM subcohort (N [ 390)
1.9
164 (15.6)
0 (0.0)
4.9
Hgb (range), g/L
12.2 (6.0-18.0)
2.1
31 (3.0)
0 (0.0)
6.0
Urea (range), mmol/L
6.4 (1.0-38.0)
2.3
39 (3.7)
0 (0.0)
7.2
Cardiac failure, no. (%)
2.5
1 (0.1)
0 (0.0)
8.6
None/mild
354 (90.8)
2.7
59 (5.6)
0 (0.0)
10.3
Moderate
27 (6.9)
2.9
10 (1.0)
0 (0.0)
12.4
Severe
9 (2.3)
3.2
1 (0.1)
0 (0.0)
16.0
3.4
1 (0.1)
0 (0.0)
18.9
3.6
3 (0.3)
0 (0.0)
22.1
ACPGBI, Association of Coloproctology of Great Britain and Ireland; ASA, American Society of Anesthesiologists; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; Hgb, hemoglobin.
ACPGBI, Association of Coloproctology of Great Britain and Ireland.
resection is still required, and comparative safety data are still largely lacking. In this large, prospective, multicenter cohort of patients with colonic AMNs treated by EMR, endoscopic mortality was 0 and significantly less than surgical mortality as predicted by 2 well-validated scoring systems. Previous studies on colorectal surgery outcomes for patients with CRC have shown that the ACPGBI score18,21,22,26 and CR-POSSUM accurately predict surgical mortality.19,21,22,27-30 Furthermore, where data were incomplete, we adopted a conservative model that favored the surgical approach by assuming normal scores for these patients, thus likely artificially reducing the predicted surgical mortality. Even so, there was a significant advantage of actual endoscopic mortality over predicted surgical mortality. We used 2 different risk scores to predict surgical mortality in our patient cohort: 1 that was designed for patients with CRC (ACPGBI score) and 1 that was designed for colorectal surgery including benign pathology (CR-POSSUM). It is important to note that both scoring systems predicted a similar mortality of 3.3% and 3.2%, respectively. As both scoring systems use quite different physiological variables apart from age, the similar predicted mortality is likely to be valid. Importantly, the actual endoscopic mortality rate was significantly lower than the predicted surgical mortality rate, even in relatively healthy patients (ASA 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
grades I and II). Even if we assume that predicted surgical mortality was overestimated, 2 large prospective case series with 750 consecutive patients undergoing laparoscopic colonic or rectal surgery31 and 15,427 patients undergoing open colonic segmental resection32 found mortality rates of 2.2% and 4.4%, respectively, both still significantly higher than the endoscopic mortality rate in our cohort.31,32 CR-POSSUM scoring was used on an unselected subcohort, which confirmed results from ACPGBI scoring. To minimize selection bias through better recall, ie, better availability of blood tests and cardiac history on more debilitated patients, we assumed lowest possible CR-POSSUM and thus lowest mortality rate for patients with missing data in the analysis of the total cohort. Thus, overall mortality per the CR-POSSUM in the whole study cohort was rather underestimated than overestimated with nevertheless a highly significant difference between actual endoscopic and predicted surgical mortality. This study was unique in several aspects. It is by far the largest, prospectively collected patient cohort resulting in good statistical power, especially as all referred patients were included without exceptions. Our study was designed as a multicenter study in 7 separate, academic tertiary centers in Australia, collecting prospectively a large group of www.giejournal.org
Ahlenstiel et al
Endoscopic mortality for EMR of colonic AMN
TABLE 5. Actual endoscopic and predicted surgical mortality No.
AEM, no. (%)
PSM, no. (%)
OR (CI)
RR (CI)
P value
NNT
All patients
1050
0 (0)
35 (3.3)
0 (0-NaN)
0.4915 (0.4704-0.5136)
!.0001
30
Successful EMR
917
0 (0)
30 (3.3)
0 (0-NaN)
0.4917 (0.4691-0.5153)
!.0001
30
All patients
390
0 (0)
13 (3.2)
0 (0-NaN)
0.4915 (0.4574-0.5282)
.0003
30
Successful EMR
336
0 (0)
11 (3.3)
0 (0-NaN)
0.4917 (0.4550-0.5313)
.0009
31
All patients
1061
0 (0)
25 (2.4)
0 (0-NaN)
0.494 (0.4731-0.5159)
!.0001
42
Successful EMR
921
0 (0)
22 (2.4)
0 (0-NaN)
0.494 (0.4715-0.5175)
!.0001
42
Group ACPGBI (complete data)
CR-POSSUM (complete data)
CR-POSSUM (total cohort)
AEM, Actual endoscopic mortality; PSM, predicted surgical mortality; OR, odds ratio; CI, confidence interval; RR, risk ratio; NNT, number needed to treat; ACPGBI, Association of Coloproctology of Great Britain and Ireland; CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity; NaN, not a number.
TABLE 6. Mortality stratified by CR-POSSUM
TABLE 7. Predicted surgical mortality by CR-POSSUM stratified by urea
No. (%)
Actual endoscopic mortality
Predicted surgical mortality, %
6
46 (11.8)
0 (0.0)
0.7
7
35 (9.0)
0 (0.0)
1.0
8
69 (17.7)
0 (0.0)
1.3
9
93 (23.8)
0 (0.0)
1.9
10
57 (14.6)
0 (0.0)
2.6
11
22 (5.6)
0 (0.0)
3.6
12
3 (0.8)
0 (0.0)
5.0
13
19 (4.9)
0 (0.0)
6.9
14
27 (6.9)
0 (0.0)
9.3
15
8 (2.1)
0 (0.0)
12.6
16
7 (1.8)
0 (0.0)
16.8
17
2 (0.5)
0 (0.0)
22.0
18
2 (0.5)
0 (0.0)
28.4
Physiological score
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity.
patients undergoing EMR for AMNs within a reasonably short time (34 months). This ensured continuity in terms of equipment standards, endoscopic procedures, and protocol, as well as consultant endoscopists performing the procedures. Also data collection was comprehensive and follow-up was complete. The best evidence to confirm a 30-day mortality advantage of EMR against surgery would undoubtedly be derived www.giejournal.org
Predicted surgical mortality
Urea, no. (%) %10
10.1-15.0
O15.0
0.7
46 (13.2)
0 (0.0)
0 (0.0)
1.0
35 (10.0)
0 (0.0)
0 (0.0)
1.3
68 (19.5)
1 (3.4)
0 (0.0)
1.9
90 (25.8)
1 (3.4)
2 (16.7)
2.6
51 (14.6)
5 (17.2)
1 (8.3)
3.6
9 (2.6)
13 (44.8)
0 (0.0)
5.0
0 (0.0)
0 (0.0)
3 (25.0)
6.9
18 (5.2)
1 (3.4)
0 (0.0)
9.3
27 (7.7)
0 (0.0)
0 (0.0)
12.6
4 (1.1)
4 (13.8)
0 (0.0)
16.8
1 (0.3)
4 (13.8)
2 (16.7)
22.0
0 (0.0)
0 (0.0)
2 (16.7)
28.4
0 (0.0)
0 (0.0)
2 (16.7)
Total
349
29
12
CR-POSSUM, Colorectal Physiologic and Operative Severity Score for Enumeration of Mortality and Morbidity.
from a multicenter, randomized, controlled trial. However, such a study would seem potentially unethical and logistically very challenging to perform. EMR has proven itself to be a safe and effective therapy for AMN in patients with a predicted 30-day mortality from elective surgery of more Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 7
Endoscopic mortality for EMR of colonic AMN
Ahlenstiel et al
than 5% or more than 10% (comprising 13.8% and 7% of our cohort, respectively, by ACPGBI) could not be ethically assigned to the surgery arm. Given the large number of prospectively recruited patients, comorbidities, and age in the study, this cohort is likely to be representative of patients needing treatment for colonic AMN. Nevertheless, it is difficult to rule out a selection bias with our patients negatively preselected toward comorbidity, ie, their comorbidities may be a reason why they were referred for EMR rather than surgery. However, 87% of the group were ASA I or II, and even in these relatively healthy patients, the 30-day mortality was significantly lower for EMR than predicted for surgery, suggesting that EMR is superior even in relatively well patients. Beyond perioperative mortality risk, both the endoscopic and the surgical approaches have their own benefit and risk profile. Surgery is usually definitive, but carries the potential of major morbidity in 9%33 and anastomotic leak in 6% to 7% with extended hospital stay.34 The main adverse event with EMR is significant postprocedural bleeding necessitating hospital admission, which occurs in 6% to 7% overall.35 Perforation is infrequent at 1% to 2% and can now be recognized and managed effectively endoscopically in the majority of patients without major sequelae.24 The major limitation of EMR is the need for scheduled colonoscopic surveillance at 4 to 6 months after EMR. There is recurrence in approximately 15% to 20%, but it is usually minor, unifocal, and easily treated. Finally, with increasing life expectancy, improving management of comorbidities worldwide, and increased uptake of CRC screening, the necessity to remove premalignant lesions in elderly patients with substantial comorbidities is likely to grow. A good example is the patient with renal disease in whom cancer is the second most common cause of death.36 Our study included a significant number of patients with renal disease (Table 7) and no mortality from EMR. Currently, the limitations of an EMR strategy for the treatment of AMN is the need for an experienced endoscopist, a dedicated team, suitable endoscopic equipment, and surgical support, a combination usually only available in tertiary centers and therefore currently not yet readily accessible everywhere. Nevertheless, our study underlines the need to establish more centers capable of performing safe and effective EMR, considering that demand is increasing with an aging population with increasing comorbidities. In summary, the data from this large multicenter cohort suggest that EMR is not only noninferior, but is safer than traditionally accepted surgical resection for the treatment of colonic AMN. Given that endoscopic therapy is less morbid and expensive than surgery and can be performed as an outpatient, endoscopic resection should be considered the first-line treatment for colonic AMNs. 8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2014
REFERENCES 1. AIHW, CA (Cancer Australia) & AAACR (Australasian Association of Cancer Registries), 2008 R. Cancer survival and prevalence in Australia: cancers diagnosed from 1982 to 2004. Cancer series no. 42. Cat. No. CAN 38. Canberra (Australia): AIHW. 2. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 1975;116:544-73. 3. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977-81. 4. Zauber AG, Winawer SJ, O'Brien MJ, et al. Colonoscopic polypectomy and long-term prevention of colorectal-cancer deaths. N Engl J Med 2012;366:687-96. 5. Soetikno RM, Kaltenbach T, Rouse RV, et al. Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults. JAMA 2008;299:1027-35. 6. Voloyiannis T, Snyder MJ, Bailey RR, et al. Management of the difficult colon polyp referred for resection: resect or rescope? Dis Colon Rectum 2008;51:292-5. 7. Onken JE, Friedman JY, Subramanian S, et al. Treatment patterns and costs associated with sessile colorectal polyps. Am J Gastroenterol 2002;97:2896-901. 8. Manfredi S, Piette C, Durand G, et al. Colonoscopy results of a French regional FOBT-based colorectal cancer screening program with high compliance. Endoscopy 2008;40:422-7. 9. McNicol L, Story DA, Leslie K, et al. Postoperative complications and mortality in older patients having non-cardiac surgery at three Melbourne teaching hospitals. Med J Aust 2007;186:447-52. 10. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128-37. 11. Swan MP, Bourke MJ, Alexander S, et al. Large refractory colonic polyps: is it time to change our practice? A prospective study of the clinical and economic impact of a tertiary referral colonic mucosal resection and polypectomy service (with videos). Gastrointest Endosc 2009;70:1128-36. 12. Moss A, Bourke MJ, Williams SJ, et al. Endoscopic mucosal resection outcomes and prediction of submucosal cancer from advanced colonic mucosal neoplasia. Gastroenterology 2011;140:1909-18. 13. Uraoka T, Saito Y, Matsuda T, et al. Endoscopic indications for endoscopic mucosal resection of laterally spreading tumours in the colorectum. Gut 2006;55:1592-7. 14. Conio M, Repici A, Demarquay JF, et al. EMR of large sessile colorectal polyps. Gastrointest Endosc 2004;60:234-41. 15. Buchner AM, Guarner-Argente C, Ginsberg GG. Outcomes of EMR of defiant colorectal lesions directed to an endoscopy referral center. Gastrointest Endosc 2012;76:255-63. 16. Tekkis PP, Poloniecki JD, Thompson MR, et al. Operative mortality in colorectal cancer: prospective national study. BMJ 2003;327:1196-201. 17. Tekkis PP, Prytherch DR, Kocher HM, et al. Development of a dedicated risk-adjustment scoring system for colorectal surgery (colorectal POSSUM). Br J Surg 2004;91:1174-82. 18. Al-Homoud S, Purkayastha S, Aziz O, et al. Evaluating operative risk in colorectal cancer surgery: ASA and POSSUM-based predictive models. Surg Oncol 2004;13:83-92. 19. Horzic M, Kopljar M, Cupurdija K, et al. Comparison of P-POSSUM and Cr-POSSUM scores in patients undergoing colorectal cancer resection. Arch Surg 2007;142:1043-8. 20. Vather R, Zargar-Shoshtari K, Adegbola S, et al. Comparison of the possum, P-POSSUM and Cr-POSSUM scoring systems as predictors of postoperative mortality in patients undergoing major colorectal surgery. ANZ J Surg 2006;76:812-6. 21. Yan J, Wang YX, Li ZP. Predictive value of the POSSUM, p-POSSUM, crPOSSUM, APACHE II and ACPGBI scoring systems in colorectal cancer resection. J Int Med Res 2011;39:1464-73.
www.giejournal.org
Ahlenstiel et al 22. Ferjani AM, Griffin D, Stallard N, et al. A newly devised scoring system for prediction of mortality in patients with colorectal cancer: a prospective study. Lancet Oncol 2007;8:317-22. 23. Senagore AJ, Warmuth AJ, Delaney CP, et al. POSSUM, p-POSSUM, and Cr-POSSUM: implementation issues in a United States health care system for prediction of outcome for colon cancer resection. Dis Colon Rectum 2004;47:1435-41. 24. Holt BA, Bourke MJ. Wide field endoscopic resection for advanced colonic mucosal neoplasia: current status and future directions. Clin Gastroenterol Hepatol 2012;10:969-79. 25. Friedland S, Banerjee S, Kochar R, et al. Outcomes of repeat colonoscopy in patients with polyps referred for surgery without biopsyproven cancer. Gastrointest Endosc 2013;79:101-7. 26. Teeuwen PH, Bremers AJ, Groenewoud JM, et al. Predictive value of POSSUM and ACPGBI scoring in mortality and morbidity of colorectal resection: a case-control study. J Gastrointest Surg 2011;15: 294-303. 27. Bromage SJ, Cunliffe WJ. Validation of the CR-POSSUM risk-adjusted scoring system for major colorectal cancer surgery in a single center. Dis Colon Rectum 2007;50:192-6. 28. Leung E, Ferjani AM, Stellard N, et al. Predicting post-operative mortality in patients undergoing colorectal surgery using P-POSSUM and CRPOSSUM scores: a prospective study. Int J Colorectal Dis 2009;24: 1459-64. 29. Leung E, Ferjani AM, Kitchen A, et al. Risk-adjusted scoring systems can predict surgeons’ performance in colorectal surgery. Surgeon 2011;9: 3-7. 30. Ptok H, Marusch F, Schmidt U, et al. Risk adjustment as basis for rational benchmarking: the example of colon carcinoma. World J Surg 2011;35:196-205.
www.giejournal.org
Endoscopic mortality for EMR of colonic AMN 31. Schlachta CM, Mamazza J, Gregoire R, et al. Could laparoscopic colon and rectal surgery become the standard of care? A review and experience with 750 procedures. Can J Surg 2003;46:432-40. 32. Prystowsky JB, Bordage G, Feinglass JM. Patient outcomes for segmental colon resection according to surgeon’s training, certification, and experience. Surgery 2002;132:663-70. 33. Cirocchi R, Farinella E, Trastulli S, et al. Elective sigmoid colectomy for diverticular disease. Laparoscopic vs open surgery: a systematic review. Colorectal Dis 2012;14:671-83. 34. Krarup PM, Jorgensen LN, Andreasen AH, et al. A nationwide study on anastomotic leakage after colonic cancer surgery. Colorectal Dis 2012;14:e661-7. 35. Burgess NG, Metz AJ, Williams SJ, et al. Risk factors for intraprocedural and clinically significant delayed Bleeding after wide-field endoscopic mucosal resection of large colonic lesions. Clin Gastroenterol Hepatol 2014;12:651-61.e3. 36. Australian and New Zealand Dialysis and Transplant Registry, The 30th Annual Report, 2007 Available at http://www.anzdata.org.au/v1/ annual_reports_download.html. Melbourne, Victoria (5), Department of Gastroenterology and Hepatology, Western Hospital, Melbourne, Victoria (6), Department of Gastroenterology and Hepatology, Lyell McEwin Hospital, Adelaide, South Australia (7), Australia. Michael J. Bourke, MBBS, FRACP, Department of Gastroenterology and Hepatology, Westmead Hospital, Hawkesbury Road, Westmead, NSW 2145, Australia. If you would like to chat with an author of this article, you may contact Dr Ahlenstiel at [email protected].
Volume
-,
No.
-
: 2014 GASTROINTESTINAL ENDOSCOPY 9