Surg Endosc DOI 10.1007/s00464-015-4579-x
and Other Interventional Techniques
Predictive factors for successful ultrasound-guided percutaneous drainage in necrotizing pancreatitis Qiang Guo1 • Ang Li1 • Weiming Hu1
Received: 21 June 2015 / Accepted: 19 September 2015 Ó Springer Science+Business Media New York 2015
Abstract Background Percutaneous catheter drainage (PCD) is now regarded as an initial minimal access technique of step-up approach for necrotizing pancreatitis. Factors that led to surgical intervention after initial management with PCD have rarely been reported. This study was to evaluate the safety and efficacy of ultrasound-guided PCD in patients with necrotizing pancreatitis and identify a subgroup of patients where PCD alone would be effective. Methods We performed a retrospective review of patients with necrotizing pancreatitis who underwent intervention in West China Hospital from January 1, 2009, to March 31, 2013. Results Patients who underwent initial PCD therapy had lower intra-abdominal bleeding rate (41/235 vs. 1/51, P = 0.002), lower enterocutaneous fistula rate (28/235 vs. 0/51, P = 0.004), and lower mortality rate (46/235 vs. 3/51, P = 0.001) when compared with the patients who underwent operative intervention. The successful PCD group had lower computed tomography (CT) mean density of necrotic fluid collection (18 HU vs. 25 HU, P = 0.01) and higher prevalence of walled-off necrosis (20/35 vs. 5/16, P = 0.04) when compared with failed PCD group. Multivariate analysis of the predictors of surgery showed that only CT mean density of necrotic fluid collection [odd ratio (OR) 1.63, 95 % confidence interval (CI) 1.04–2.94, P = 0.006] was identified as significant factor.
Qiang Guo and Ang Li have contributed equally to this study. & Weiming Hu
[email protected] 1
Pancreatic Surgery, West China Hospital, Sichuan University, Chengdu 610041, Sichuan Province, China
Conclusion CT mean density of necrotic fluid collection and the existence of acute necrotic collection could influence the success rate of PCD. Keywords Acute pancreatitis Percutaneous catheter drainage Operative intervention Pancreatic necrosis
Most patients with acute pancreatitis develop a mild course and recover without intervention in a short, uncomplicated hospitalization [1, 2]. The remaining patients develop a severe course that requires long hospitalization with a high mortality rate [1, 2]. Some of the patients with severe acute pancreatitis need intensive care and invasive intervention [3]. Infection or suspected infection of pancreatic necrosis is well-accepted indications for intervention [4]. Operative necrosectomy is associated with high morbidity and mortality because it is always a second hit to the critical condition of the patient [5]. Many minimally invasive techniques such as percutaneous catheter drainage (PCD) [6], endoscopic necrosectomy [7], and laparoscopic pancreatic necrosectomy were explored to decrease the mortality and morbidity [8]. PCD is now regarded as the first choice for necrotizing pancreatitis [9]. It is proved to be successful in 30–90 % patients and associated with lower complication rate when compared with open necrosectomy [10–15]. PCD is always performed with CT guidance, whereas ultrasound-guided PCD has few mentioned [16, 17]. Factors that led to surgical intervention after initial management with PCD have also rarely been reported [18, 19]. The aim of this study was to evaluate the safety and efficacy of ultrasoundguided PCD in patients with necrotizing pancreatitis and identify a subgroup of patients where PCD alone would be effective.
123
Surg Endosc
Materials and methods This retrospective study was performed by searching the medical records of the patients with necrotizing pancreatitis who underwent intervention hospitalized at West China Hospital from January 1, 2009, to March 31, 2013. The data collected for each patient included demographic data (age, sex, and etiology), clinical information [Acute Physiology and Chronic Health Evaluation II score (APACHE II score), Balthazar CT score, mean CT density and feature of necrosis, C-reactive protein value, timing of intervention, organ failure, the status of necrosis], and clinical outcome (new-onset organ failure, intra-abdominal bleeding, enterocutaneous fistula, and mortality). The process of patient selection and therapeutic management in this series is illustrated in Fig. 1. The feature of necrosis referred to acute necrotic collection or walled-off necrosis (WON) according to Revised Atlanta Classification [20]. The mean CT density of fluid collection was measured by the last CT scan before PCD using MDC PACS Viewer system (Philip Corporation, Netherlands). The status of necrosis referred to sterile or infected necrosis confirmed by the bacterial culture of fluid collection obtained during intervention. Intra-abdominal bleeding was defined as bleeding requiring surgical, radiological, or endoscopic intervention. Enterocutaneous fistula was defined as secretion of fecal material from a percutaneous drain or drainage canal after the removal of drains or from a surgical wound; secretion comes from either the small or large bowel and is confirmed with imaging or during surgery and requires intervention. New-
Fig. 1 Study population
123
onset organ failure was defined as organ failure that had not been present before intervention. All patients underwent non-operative treatment (including antibiotics, nutrition support, and organ function maintenance) initially. We only performed primary PCD on the patients with fluid collection that was close to the abdominal wall and, at the same time, the extent of pancreatic necrosis less than 30 % based on contrast-enhanced computed tomography (Fig. 2). Operative intervention was performed on the remaining patients. We conducted this selective PCD strategy because we believe that if the fluid collection was not close to the abdominal wall, it is easy to result in bleeding and gastrointestinal perforation. Besides, large amount of pancreatic necrosis might lead to obstruction of the tube and repeated procedures. All drainages were performed under color ultrasound using 12-Fr or 14-Fr pigtail tube. The tube was placed as far as possible into the fluid collection to ensure effective drainage. Irrigation was started the next day after intervention with 0.9 % saline solution twice a day. If no clinical improvement (persistence or recurrence of clinical manifestations of sepsis), inadequate drainage, or the presence of severe complications were observed, operative necrosectomy was performed (Fig. 3). Characteristics and outcomes were compared between the operative necrosectomy group and the PCD group. Characteristics and outcomes the successful PCD group and the failed PCD group were also compared. The data analysis was made using SPSS 17.0 statistical software (Statistical Product and Service Solutions, Chicago, USA). Continuous data are evaluated using t test; categorical data are described with frequency counts and percentages and
Surg Endosc
Fig. 2 CECT scan of a 46-year-old male patient with acute pancreatitis 28 days after onset of symptoms. Normal enhancement of the pancreas (white asterisks) existed. A single fluid collection was close to abdominal wall. The black arrow points at the direction of puncture
Fig. 3 CECT scan of a 48-year-old male patient 8 days after an initial PCD. Gas bubbles were seen in the necrotic collection, and persistence of manifestations of infection existed
assessed using Chi-square test. Multivariate analysis was used to determine the effect of the different predictors of surgery. Two-sided P \ 0.05 was considered statistically significant.
Results During the study period, a total of 286 patients underwent intervention. Of these 286 patients, 174 were confirmed to have infected necrosis. The most common cause of pancreatitis was biliary (142/286, 50 %). The overall mortality rate was 17 % (49/286).
Fifty-one patients were treated with PCD therapy initially. The others underwent operative intervention. Characteristics and outcomes of these two groups are listed in Table 1. The demographic data were similar in the two groups. However, as we described above, because of the selection criteria, the PCD group had lower CT severity index than the surgery group. The PCD group also had lower APACHE II score on admission and lower prevalence of multiple organ failure before intervention (66/235 vs. 7/51, P = 0.03). With respect to the outcomes, the PCD group had lower intra-abdominal bleeding rate (41/235 vs. 1/51, P = 0.002), lower enterocutaneous fistula rate (28/235 vs. 0/51, P = 0.004), and lower mortality rate (46/235 vs. 3/51, P = 0.001) when compared with the surgery group. Thirty-five patients were successfully managed by PCD alone, making the success rate 69 %. The others were converted to surgery. There was no statistically significant difference in most demographic and clinical characteristics between the successful PCD group and the failed PCD group (Table 2). However, CT features of these two groups were different. The successful PCD group had lower CT mean density of necrotic fluid collection (18 HU vs. 25 HU, P = 0.01) and higher prevalence of walled-off necrosis (20/35 vs. 5/16, P = 0.04) when compared with failed PCD group. Fifteen of 16 patients in the failed PCD group experienced persistence or recurrence of clinical manifestations of sepsis. Only one patient converted to surgery because of intra-abdominal bleeding. Organ failure before intervention, infected necrosis, pancreatic necrosis, CT mean density of necrotic fluid collection, and walledoff necrosis were entered into a multivariate analysis of the predictors of surgery. Only CT mean density of necrotic fluid collection (OR 1.63, 95 % CI 1.04–2.94, P = 0.006) was identified as significant factor.
Discussion In 1998, Freeny et al. [6] first described the use of imagingguided PCD as an alternative treatment to the patients with infected pancreatic necrosis. During the past decade, several minimally invasive procedures, such as video-assisted retroperitoneal debridement, endoscopic transgastric necrosectomy, and transluminal endoscopic necrosectomy, were introduced [21]. Results of many trails have revealed that PCD was a safe and efficient therapy in the treatment of patients with necrotizing pancreatitis. We started to perform ultrasound-guided PCD since 2009. The selection criteria of this procedure were based on contrast-enhanced computed tomography features: (a) fluid collection that was close to the abdominal wall and (b) the extent of pancreatic necrosis less than 30 %. This retrospective study investigated the effectiveness and safety of the use of PCD
123
Surg Endosc Table 1 Clinical characteristics and outcomes of the patients received intervention Characteristic
Surgery (N = 235)
Age, median (range) (years)
47 (20–75)
Male [no. (%)]
PCD (N = 51)
P value
48 (22–76)
0.35
143 (61)
32 (63)
0.88
Etiology [no. (%)] Biliary
123 (52)
19 (37)
0.06
Alcohol abuse
24 (10)
7 (14)
0.46
Others
88 (38)
25 (49)
APACHE II score on admission
0.16
8 (2–30)
6 (2–25)
\0.001
8 (4–10)
5 (3–6)
\0.001
Computed tomography CT severity index, median (range), HU Pancreatic necrosis [no. (%)] Infected pancreatic necrosis [no. (%)] Organ failure Organ failure at any time before intervention [no. (%)] Multiple organ failure at any time before intervention [no. (%)] C-reactive protein at admission, median (range) (mg/L)
167 (71)
31 (61)
0.18
146 (62)
28 (55)
0.35
108 (46)
18 (35)
0.17
66 (28)
7 (14)
0.03
210 (34–433)
202 (44–398)
Timing of intervention, median (range) (days)
33 (7–95)
30 (13–43)
New-onset organ failure after first intervention (no. (%)]
33 (14)
4 (8)
0.46 0.12 0.26
Intra-abdominal bleeding [no. (%)]
41 (17)
1 (2)
0.002
Enterocutaneous fistula [no. (%)]
28 (12)
0 (0)
0.004
Mortality [no. (%)]
46 (20)
3 (6)
0.01
Table 2 Clinical characteristics and outcomes of the patients received PCD Characteristic
Successful PCD (N = 35)
Failed PCD (N = 16)
P value
Age, median (range) (years)
47 (22–70)
49 (26–76)
0.35
Male [no. (%)]
20 (57)
12 (75)
0.35
11 (31)
8 (50)
0.23
4 (11)
3 (19)
0.66
20 (58)
5 (31)
0.13
6 (2–23)
6 (2–25)
0.55
5 (3–6)
5 (3–6)
0.65
8 (50)
0.36
Etiology [no. (%)] Biliary Alcohol abuse Others APACHE II score on admission Computed tomography CT severity index, median (range) Pancreatic necrosis [no. (%)]
23 (66)
CT mean density of fluid collection [no. (%)] B30 HU
32 (91)
5 (31)
\0.001
3 (9)
11 (69)
\0.001
Walled-off necrosis [no. (%)]
20 (57)
4 (25)
0.04
Acute necrotic collection [no. (%)]
15 (43)
12 (75)
0.04
Infected pancreatic necrosis [no. (%)]
18 (51)
10 (63)
0.55
11 (31)
7 (44)
0.53
4 (11)
3 (19)
[30 HU
Organ failure Organ failure at any time before intervention [no. (%)] Multiple organ failure at any time before intervention [no. (%)] C-reactive protein at admission, median (range) (mg/L) Timing of intervention, median (range) (days) Mortality [no. (%)]
123
196 (44–398)
212 (64–376)
31 (16–43)
28 (13–34)
0 (0)
3 (6)
0.66 0.45 0.17 0.03
Surg Endosc
in patients with necrotizing pancreatitis. Our results indicate that ultrasound-guided PCD, as compared with operative intervention, has a lower mortality rate and complication rate. By comparing the characteristics and outcomes of successful PCD and failed PCD and conducting a multivariate analysis, we confirmed that CT mean density of necrotic fluid collection was a significant predictor of surgery. The PANTER (Pancreatitis, Necrosectomy versus StepUp Approach) trial was conducted to compare the step-up approach, consisting of PCD as first step followed by video-assisted retroperitoneal debridement if necessary, with the traditional procedure, primary open necrosectomy [9]. Results showed that of the patients assigned to the stepup approach, 35 % were successfully treated with ultrasound-/CT-guided PCD only. New-onset multiple organ failure occurred less often in patients assigned to the stepup approach than in those assigned to open necrosectomy; however, the rate of death did not differ significantly between groups. Another cohort study also compared these two procedures. Differed from the PANTER trial, PCD in this study was applied under the guidance of ultrasound and the mortality rate was lower in the PCD group than in the surgery group. In our study, the selective ultrasoundguided PCD procedure had lower intra-abdominal bleeding rate (41/235 vs. 1/51, P = 0.002), lower enterocutaneous fistula rate (28/235 vs. 0/51, P = 0.004), and lower mortality rate (46/235 vs. 3/51, P = 0.001) when compared with the operative intervention procedure. All these results support the hypothesis that PCD has less surgical trauma in the patients who are already severely ill, which is the most important advantage of PCD. However, it should be mentioned that, because of the selection criteria, the PCD group had lower CT severity index than the surgery group. The PCD group also had lower APACHE II score on admission and lower prevalence of multiple organ failure before intervention. These factors might impact on the outcomes of the two groups. Despite the excellent efficacy of PCD, it should be mentioned that based on the results of a systemic review, about 5–70 % of the patients need additional surgery [22]. Persistent sepsis, inadequate drainage, or the presence of severe complications (such as bleeding or enterocutaneous fistula) could significantly influence the success rate of PCD [17–19]. In our study, 31 % of the patients who underwent initial PCD had additional surgery. Considering the reasons, 15 of 16 patients in the failed PCD group experienced persistence or recurrence of clinical manifestations of sepsis and only one patient converted to surgery because of intra-abdominal bleeding. There are two previous studies focused on predictive factors that led to surgical intervention after initial management with PCD [18, 19]. Babu et al. [18] described a
prospective study including 70 patients, 14 of which were managed medically, 29 managed with PCD alone, whereas 27 required surgery after initial PCD. They concluded that APACHE II score at first intervention and organ failure within a week of the onset of disease could predict the need for surgery in the early course of disease. Unfortunately, the authors did not provide any explanation about why organ failure would lead to failure of PCD. Tong et al. [19] did a retrospective review of 34 patients with infective pancreatic necrosis. Nineteen of 34 patients were cured by PCD alone, whereas open necrosectomy were needed for 15 patients. The mean computed tomographic density and distribution range of infective pancreatic necrosis were identified as predictive factors of surgery after initial PCD, which were not mentioned in Babu’s study. In our study, the successful PCD group had lower CT mean density of necrotic fluid collection and higher prevalence of walledoff necrosis (20/35 vs. 5/16, P = 0.04) when compared with failed PCD group. After a multivariate analysis of the possible predictors of surgery, only CT mean density of necrotic fluid collection (OR 1.63, 95 % CI 1.04–2.94, P = 0.006) was identified as significant factor. The possible explanation was that higher CT density means a greater proportion of solid form in the necrotic fluid collection, which would lead to obstruction of drainage tube. Besides, differed from acute necrotic collection, walled-off necrosis is a mature, encapsulated collection of pancreatic and/or peripancreatic necrosis and has a well-defined inflammatory wall. Thus, fluid collection might disperse after an initial PCD in acute necrotic collection rather than walled-off necrosis which can also lead to a failure of PCD. Most of the studies concentrated on PCD were applied under the guidance of CT, and ultrasound-guided PCD was few mentioned. Compared with CT guidance, ultrasoundguided PCD has an advantage of avoiding vessels during puncture; however, a part of the patients cannot have access for PCD under ultrasound [16, 17]. This selective ultrasound-guided PCD procedure we introduced had only one patient with intra-abdominal bleeding and no enterocutaneous fistula, lower than any other study. It might because we performed primary PCD only on the patients with fluid collection that was close to the abdominal wall. In addition, successful rate of PCD in our study reached to 69 %, higher than most studies. The disadvantage of this strategy was that only about 20 % of the patient met the selection criteria. Since studies failed to find a correlation between extent of pancreatic necrosis and failure of PCD, we will alter our PCD strategy regardless of the extent of pancreatic necrosis in the future. Recently, a matched cohort study compared clinical outcome and healthcare utilization of direct endoscopic necrosectomy versus step-up approach for walled-off pancreatic necrosis [23]. Results showed that direct
123
Surg Endosc
endoscopic necrosectomy had less new antibiotic use, pulmonary failure, endocrine insufficiency, shorter length of stay, and less healthcare utilization than step-up approach. Thus, to decrease the failure rate of PCD, it is important to select appropriate patients with necrotizing pancreatitis for PCD. Future studies should pay more attention to this issue. In conclusion, this study indicated that ultrasound-guided PCD, as compared with operative intervention, has a lower mortality rate and complication rate. Higher mean CT density of necrotic fluid collection and acute necrotic collection were associated with a higher possibility of operative intervention. Acknowledgments The corresponding author had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Author contributions All of the authors contributed to the collection and analysis of the data and to the preparation of the report.
10.
11.
12.
13.
14.
Compliance with ethical standards Disclosures Drs. Qiang Guo, Ang Li, and Weiming Hu have no conflicts of interest or financial ties to disclose.
15.
References 16. 1. Acevedo-Piedra NG, Moya-Hoyo N, Rey-Riveiro M, Gil S, Sempere L, Martı´nez J, Lluı´s F, Sa´nchez-Paya´ J, de-Madaria E (2014) Validation of the determinant-based classification and revision of the Atlanta classification systems for acute pancreatitis. Clin Gastroenterol Hepatol 12:311–316 2. Nawaz H, Mounzer R, Yadav D, Yabes JG, Slivka A, Whitcomb DC, Papachristou GI (2013) Revised Atlanta and determinantbased classification: application in a prospective cohort of acute pancreatitis patients. Am J Gastroenterol 108:1911–1917 3. Banks PA, Freeman ML (2006) Practice guidelines in acute pancreatitis. Am J Gastroenterol 101:2379–2400 4. Working Group IAP/APA Acute Pancreatitis Guidelines (2013) IAP/APA evidence-based guidelines for the management of acute pancreatitis. Pancreatology 13(4 Suppl 2):1–15 5. Connor S, Alexakis N, Raraty MG, Ghaneh P, Evans J, Hughes M, Garvey CJ, Sutton R, Neoptolemos JP (2005) Early and late complications after pancreatic necrosectomy. Surgery 137: 499–505 6. Freeny PC, Hauptmann E, Althaus SJ, Traverso LW, Sinanan M (1998) Percutaneous CT-guided catheter drainage of infected acute necrotizing pancreatitis: techniques and results. AJR Am J Roentgenol 170:969–975 7. Papachristou GI, Takahashi N, Chahal P, Sarr MG, Baron TH (2007) Peroral endoscopic drainage/debridement of walled-off pancreatic necrosis. Ann Surg 245:943–951 8. Horvath KD, Kao LS, Ali A, Wherry KL, Pellegrini CA, Sinanan MN (2001) Laparoscopic assisted percutaneous drainage of infected pancreatic necrosis. Surg Endosc 15:677–682 9. van Santvoort HC, Besselink MG, Bakker OJ, Hofker HS, Boermeester MA, Dejong CH, van Goor H, Schaapherder AF,
123
17.
18.
19.
20.
21. 22.
23.
van Eijck CH, Bollen TL, van Ramshorst B, Nieuwenhuijs VB, Timmer R, Lame´ris JS, Kruyt PM, Manusama ER, van der Harst E, van der Schelling GP, Karsten T, Hesselink EJ, van Laarhoven CJ, Rosman C, Bosscha K, de Wit RJ, Houdijk AP, van Leeuwen MS, Buskens E, Gooszen HG, Dutch Pancreatitis Study Group (2010) A step-up approach or open necrosectomy for necrotizing pancreatitis. N Engl J Med 362:1491–1502 Gambiez LP, Denimal FA, Porte HL, Saudemont A, Chambon JP, Quandalle PA (1998) Retroperitoneal approach and endoscopic management of peripancreatic necrosis collections. Arch Surg 133:66–72 Fotoohi M, D’Agostino HB, Wollman B, Chon K, Shahrokni S, vanSonnenberg E (1999) Persistent pancreatocutaneous fistula after percutaneous drainage of pancreatic fluid collections: role of cause and severity of pancreatitis. Radiology 213:573–578 Baril NB, Ralls PW, Wren SM et al (2000) Does an infected peripancreatic fluid collection or abscess mandate operation? Ann Surg 231:361–367 Mortele´ KJ, Girshman J, Szejnfeld D, Selby RR, Radin R, Parekh D, Jabbour N, Stain SC (2009) CT-guided percutaneous catheter drainage of acute necrotizing pancreatitis: clinical experience and observations in patients with sterile and infected necrosis. AJR Am J Roentgenol 192:110–116 Rocha FG, Benoit E, Zinner MJ, Whang EE, Banks PA, Ashley SW, Mortele KJ (2009) Impact of radiologic intervention on mortality in necrotizing pancreatitis: the role of organ failure. Arch Surg 144:261–265 Baudin G, Chassang M, Gelsi E, Novellas S, Bernardin G, He´buterne X, Chevallier P (2012) CT-guided percutaneous catheter drainage of acute infectious necrotizing pancreatitis: assessment of effectiveness and safety. AJR Am J Roentgenol 199:192–199 Wron´ski M, Cebulski W, Karkocha D, Słodkowski M, Wysocki L, Jankowski M, Krasnode˛bski IW (2013) Ultrasound-guided percutaneous drainage of infected pancreatic necrosis. Surg Endosc 27:2841–2848 Ai X, Qian X, Pan W, Xu J, Hu W, Terai T, Sato N, Watanabe S (2010) Ultrasound-guided percutaneous drainage may decrease the mortality of severe acute pancreatitis. J Gastroenterol 45: 77–85 Babu RY, Gupta R, Kang M, Bhasin DK, Rana SS, Singh R (2013) Predictors of surgery in patients with severe acute pancreatitis managed by the step-up approach. Ann Surg 257: 737–750 Tong Z, Li W, Yu W, Geng Y, Ke L, Nie Y, Sun J, Ni H, Wang X, Ye X, Li N, Li J (2012) Percutaneous catheter drainage for infective pancreatic necrosis: is it always the first choice for all patients? Pancreas 41:302–305 Banks PA, Bollen TL, Dervenis C, Gooszen HG, Johnson CD, Sarr MG, Tsiotos GG, Vege SS, Acute Pancreatitis Classification Working Group (2013) Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by international consensus. Gut 62:102–111 Navaneethan U, Vege SS, Chari ST, Baron TH (2009) Minimally invasive techniques in pancreatic necrosis. Pancreas 38:867–875 van Baal MC, van Santvoort HC, Bollen TL, Bakker OJ, Besselink MG, Gooszen HG, Dutch Pancreatitis Study Group (2011) Systematic review of percutaneous catheter drainage as primary treatment for necrotizing pancreatitis. Br J Surg 98:18–27 Kumar N, Conwell DL, Thompson CC (2014) Direct endoscopic necrosectomy versus step-up approach for walled-off pancreatic necrosis: comparison of clinical outcome and health care utilization. Pancreas 43:1334–1339