Accepted Article
Prevalence and risk factors for clinically significant upper gastrointestinal bleeding in patients with severe acute pancreatitis1
Xian-Bao Zhan, Xiao-Rong Guo ,Jing Yang, Jie Li, Zhao-Shen Li
Department of Gastroenterology, Changhai Hospital, Second Military Medical University, Shanghai 200433, China
Correspondence to: Dr. Zhaoshen Li, Professor of Medicine, Chief of Department of Gastroenterology, Changhai Hospital, Second Military Medical University, Shanghai, 200433, China E-mail: [email protected]
Financial disclosure and conflict of interest: None of the authors have commercial relationships that might pose a conflict of interest in connection with this manuscript.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between
this version and the Version of Record. Please cite this article as doi: 10.1111/1751-2980.12206 1
Accepted Article
Abstract
Goals: To explore the prevalence and risk factors of upper gastrointestinal bleeding(UGIB) in patients with severe acute pancreatitis(SAP). Background: The treatment strategy of SAP changed greatly during the past decade,
but the prevalence and risk factors of the complicated UGIB are still unclear. Study: From July 2006 to June 2010, retrospective data on 101 patients with SAP admitted to Changhai Hospital were collected. The prevalence and risk factors of UGIB were analyzed. Results: Eighteen patients(17.8%) developed UGIB. Thirteen patients received
endosopic examinations which yielded 6 cases of acute gastric mucosal lesions(AGML), and 5 cases of peptic ulcer. The mortality rate of UGIB patients was much higher than that in non-UGIB patients (44.4% vs 10.8%, P=0.0021). Univariate analysis revealed that the risk factors for UGIB included acute physiology and chronic health evaluation II (APACHE II) score, computed tomography severity index, Ranson score, arterial blood pH values and oxygen pressures (PaO2), serum blood urea nitrogen
and
creatinine
concentrations,
prothrombin
and
activated
partial
thromboplastin times, platelet counts, the presence of shock, sepsis and organ failure other than UGIB, mechanical ventilation, heparinized continuous renal replacement therapy and total parental nutrition. By multivariate logistic regression analysis, APACHE II scores and CT severity indices(CTSI) were found to be significant risk factors, while PaO2 was the only protective factor. Conclusions: UGIB is a common complication with a poor prognosis due mainly to
2
Accepted Article
peptic ulcer and AGML. Patients with high APACHE II scores, and CTSI or low PaO2 should be considered to be at high risk for UGIB.
Key words: acute pancreatitis;upper gastrointestinal bleeding;risk factor;peptic ulcer;acute gastric mucosal lesion
3
Accepted Article
1. Introduction 2. Severe acute pancreatitis (SAP) occurs in approximately 20% of cases of acute pancreatitis, and is associated with a 15% to 30% mortality rate [1]. The outcome of SAP mostly depends on whether systemic and local complications develop. It has been recognized that upper gastrointestinal bleeding (UGIB) is one of the most common complications with SAP [2-4]. Previous reports have shown that 50–65% of patients with acute pancreatitis may have acute gastrointestinal mucosal lesions [2-3]. According to the definition established at the Atlanta Symposium in 1992 and UK guidelines for the management of acute pancreatitis in 2005, gastrointestinal bleeding greater than 500 mL in 24 h can be diagnosed as gastrointestinal failure [5,6]. During the last decade, the treatment of SAP has
changed greatly. Currently, conservative therapy is the treatment of choice for SAP [7-8]. Although the prophylactic use of proton pump inhibitors been widely adopted in SAP treatment regimens, the current prevalence and risk factors of UGIB in patients with SAP treated prophylactically with proton pump inhibitors is
unclear. This study aimed to shed light on this issue.
4
Accepted Article
2. Materials and Methods
We conducted a retrospective chart review of patients admitted to the specialized pancreatic intensive care unit at Changhai Hospital between July 2006 and June 2010. Changhai Hospital is a tertiary and academic hospital, affiliated to Second Military
Medical University, and serves patients from various parts of China.
2.1 Data Collection
Charts were studied and data was retrospectively collected using a standardized data collection form. Collected variables included patient demographics, medical history, etiology of acute pancreatitis, laboratory tests (white blood cell count, red blood cell count, platelet count, hemoglobin, serum liver tests, serum creatinine and blood urea
nitrogen, prothrombin time, arterial blood gas analysis, etc.), systemic and local complications (systemic inflammatory response syndrome (SIRS), organ failure, acute fluid collection, pseudocyst, pancreatic abscess), special treatment (continuous blood purification, peritoneal lavage, mechanical ventilation, total parental nutrition (TPN), steroids use, etc.) and primary outcomes (transfer to surgery, recovery or death). The
definitions of systemic and local complications were based on the clinical classification system for acute pancreatitis established at the Altanta Symposium in 1992 and UK guidelines for the management of acute pancreatitis in 2005 [5,6].
5
Accepted Article
2.2 Patient Selection
The enrolled SAP patients were all admitted to the specialized pancreatic intensive care unit at Changhai Hospital from July 2006 to June 2010. The diagnosis of acute pancreatitis was made on clinical criteria which included elevated serum amylase (>3 times of upper limit of normal), and evidence of pancreatitis on ultrasonography or computed tomography (CT). SAP was defined as an (APACHE II) of 8 or more, or Ranson score of 3 or more [1, 6-7]. All patients with SAP underwent detailed clinical
evaluation and investigation, including monitoring of vital signs, hematology and serum biochemistry tests, arterial blood gas analysis, and plain film radiography of the abdomen and chest. A dynamic contrast-enhanced CT was done around 72 h after the onset of SAP. During the first 72 h, active fluid resuscitation with colloids was performed. During the acute phase, usually within five to seven days after onset, proton pump inhibitors, somatostatin, gabexate and wide spectrum antibiotics were routinely used. ERCP and nasal-biliary drainage were conducted within the 72 h if acute biliary pancreatitis was diagnosed or if biliary infection existed any time. If the systemic inflammatory response syndrome existed even after conservative therapy, or if the diagnosis of hypertriglyceridemia was made, continuous renal replacement therapy (CRRT) was performed. Clinically-significant bleeding (CSB) was defined by the presence of coffee
ground gastric aspirate, gross hematemesis, or melena, with or without circulatory effects.
The
etiologies
were
classified
6
into
four
major
groups,
biliary,
Accepted Article
hypertriglyceridemia, alcoholic and other etiologies [9]. Acute biliary pancreatitis was
defined as serum alanine aminotransferase elevations >3 times of the upper limit of the normal value with or without elevations of liver function tests. Acute alcoholic pancreatitis was defined as acute pancreatitis that developed in the context of an oral intake of more than 20 g of alcohol during the past 24 hours. Acute hypertriglyceridemia-related pancreatitis was defined as acute pancreatitis that developed in the presence of a plasma triglyceride concentration over 5.65 mmol/L.
2.3. Statistical Analysis
All analyses were performed using SAS9.1 software. Frequencies, percentages, means, standard deviations, odds ratios, and confidence intervals were calculated. Continuous data were expressed as mean ± SE (standard error of the mean) determined using the Student’s t-test, while categorical variables were expressed as quantities, and were analyzed using the χ2 test. Multiple stepwise logistic regression analysis was used to identify the risk factors for UGIB. Differences in means were conducted using the
2-tailed Student’s t-test, and differences in percentages were assessed using the χ2 test. Continuous variables were tested using logistic regression methods. P values of 0.05). The median interval between
the onset of SAP and UGIB was 6 days. The mortality rate of the UGIB group was much higher than that of the non-UGIB group (P=0.0021) (Table 1).
3.2 Causes of UGIB Of the 18 patients in the UGIB group, 13 cases received upper gastrointestinal
endoscopic examinations during the hospital stay period which revealed 2 cases of gastric ulcer, 6 cases of acute gastric mucosal lesions, 3 cases of duodenal ulcer, and 2 cases of pancreatic necrotic tissue invading into the duodenal bulb presenting as a multi-lesion, honeycomb-like ulcer (Figure 1).
3.3 Risk Factors of UGIB in Patients with SAP On univariate analysis, APACHE II score, CT severity index, Ranson score, arterial blood pH value and PaO2, serum blood urea nitrogen and creatinine, prothrombin time and activated partial thromboplastin time, platelet counting, shock, sepsis, organ failure other than UGIB, mechanical ventilation time, duration of heparinized CRRT, duration of TPN, were found to be significantly associated with UGIB in patients with acute
8
Accepted Article
pancreatitis (P0.05) (Table 2). On multivariate logistic regression analysis, only the APACHE II score and CT
severity index were found to be significant risk factors, while arterial blood PaO2 was
the only protective factor (Table 3).
9
Accepted Article
4. Discussion SAP can deteriorate very rapidly. UGIB, one of the common complications of SAP, may be lethal and complicate the course of SAP. In the current study, the mortality rate of UGIB patients was found to be much higher than that in non-UGIB patients (44.4% vs. 10.8%, P=0.0021). The exact incidence of UGIB in SAP is currently unknown.
Sharma et al. reported that gastrointestinal hemorrhage occurred in 3.6% of patients
with acute pancreatitis [5]. In the current study, we found that UGIB occurred in 17.82% of patients with SAP, which is much higher than that reported by Sharma et al. The reason for the difference in incidence is likely due to that fact that the severity of
disease of our patients was greater than that reported by Sharma et al. The APACHE II score in our group was 17.61±6.02 in the UGIB group, and 9.35±6.67 in the non-UGIB group, while it was 8.7±3.2 in the gastrointestinal hemorrhage group and 5.3±5.2 in the
non-gastrointestinal hemorrhage group reported by Sharma et al. In patients with SAP, a variety of causes of UGIB have been reported including
acute gastric mucosal lesions and stress ulcer, peptic ulcer, gastric variceal bleeding,
pseudoaneurysm rupture into the duodenum or into the pancreatic duct, nasogastric tube damage, etc. [11-12]. Acute gastric mucosal lesions in patients with SAP can range from simple gastritis and erosions to ulceration or bleeding. It has been reported that the majority of the lesions or ulcers were located in the stomach and duodenum, while lesions in the esophagus were uncommon [3]. In our study, thirteen patients with SAP and UGIB received endoscopic examinations which showed 6 cases of acute
gastric mucosal lesions and 5 cases of gastric ulcer and duodenal ulcer. The results
10
Accepted Article
were consistent with those of a previous report [3]. Stress of severe illness, hypoxia,
ventilation and non-steroidal anti-inflammatory drugs for pain relief, and gastrointestinal hypoperfusion have been proposed to be potential causes of gastrointestinal mucosal barrier failure, and to contribute to the development of acute
mucosal lesions or ulcers [11, 13-14]. Animal studies have shown that prior restraint
stress aggravated pancreatic pathology from biliary-pancreatic duct outlet exclusion in closed duodenal loop models. Acute pancreatitis has been shown to exacerbate gastric damage caused by restraint stress resulting in ulceration evolving to ulcer, hemorrhage and gastric necrosis [15]. Gastrointestinal hemorrhages secondary to hemosuccus
pancreaticus or ruptured artery pseudoaneurysm are rare conditions that pose significant diagnostic and therapeutic challenges. There are a number of case reports related to such conditions [16-20]. They have been reported to occur most commonly in a setting of acute pancreatitis with rupture of pseudoaneurysms of the splenic or hepatic artery into the pancreatic duct, which were definitively diagnosed by selective celiac angiography [16]. Severe gastroduodenal necrosis with perforation has rarely
been reported [21-22]. Kubo et al. [19] described a patient with acute pancreatitis who presented with hematemesis and gastric perforation caused by gastric erosion and rupture of a pseudoaneurysm. In the current study, we found 2 cases of duodenal necrosis with perforation with refractory bleeding which resulted from erosion of pancreatic necrosis into the duodenal wall. It is difficult for clinicians to recognize this condition. A definitive diagnosis can be established by endoscopic examination. There are many vascular structures neighboring the duodenal segment, and erosion of these
11
Accepted Article
vessels may result in fatal hemorrhage. Because of the severe consequences,
physicians should be alert and consider it in the differential diagnosis. The risk factors for hemorrhagic complications of UGIB in critically ill patients
have been previously studied [23-24]. It has been proposed that old age, trauma, surgery, shock, infection, sepsis, severe jaundice, multiorgan dysfunction syndrome (MODS) or multiple systemic organ failure (MSOF), coagulation disorders, a history of ulcers during the year, long-term fasting and nutritional deficiencies were risk factors for UGIB. Although SAP is also a critical illness, there are very few reports on the risk factors for UGIB in patients with SAP within the last two decades. In the report by Sharma et al., necrotizing pancreatitis, high APACHE II score, organ failure, sepsis and pancreatic collection were significant risk factors for developing hemorrhage. However, that study included both gastrointestinal and abdominal
hemorrhages. Hagiwara et al. performed a prospective cohort study on patients with SAP to determine prognostic factors for arterial injury and gastrointestinal perforation in patients with SAP, and found that CRP ≥15 mg/dl on day 7, and a CT severity index ≥7 between days 6 and 8 were independent risk factors (P = 0.02 and 0.04, respectively). The odds ratios for CRP ≥15 mg/dl were 23.0 and 15.7, respectively for a CT severity index of ≥7. The results suggest that a persistent elevation of the CRP concentration and a high CT severity index are independent risk factors for local complications associated with SAP [16]. The current study was the first specifially designed to explore the risk factors of UGIB in patients with SAP. In the current study, we found the occurrence of UGIB had no direct relationship to the etiology of SAP,
12
Accepted Article
which was consistent with a previous report [5]. The conditions related to the risk factors in the current study actually reflect the
severity of the disease, which may have induced the gastroduodenal mucosal lesions by destroying the mucosal barrier or systemic coagulopathy. This hypothesis needs to be confirmed by further basic research. Unexpectedly, the current study found that the
liver function tests and age were not statistically different between the UGIB group and the non-UGIB group, which is not consistent with the findings of previous studies on the risk factors of stress ulcer bleeding in critically ill patients which included a variety of diseases. The reason for this needs further study, but one possible explanation might
be that the transient abnormal liver function indices do not reflect the degree of liver function loss under the conditions of acute biliary pancreatitis in the current study. The potential limitations of the current study are that it involved only a single
center, and was retrospective with a small number of patients. However, all the enrolled patients in this study were treated by the same medical group, which helped ensure that all the enrolled patients underwent similar basic management, and also allowed control of the study quality.
In conclusion, UGIB is an early common complication in patients with severe acute pancreatitis and is associated with a high mortality rate. The patients with a high APACHE II score and a CT severity index or low PaO2 should be considered at high risk for UGIB.
13
Accepted Article
References 1. Banks PA, Freeman ML. Practice guidelines in acute pancreatitis. Am J Gastroenterol 2006; 101: 2379-2400
2. Chen TA, Lo GH, Lin CK, Lai KH, Wong HY, Yu HC, Hsu PI, Chen HH, Tsai WL, Chen WC. Acute pancreatitis-associated acute gastrointestinal mucosal lesions: incidence, characteristics, and clinical significance. J Clin Gastroenterol 2007; 41:630-634
3. Lin CK, Wang ZS, Lai KH, Lo GH, Hsu PI. Gastrointestinal mucosal lesions in patients with acute pancreatitis. Zhonghua Yi Xue Za Zhi (Taipei). 2002;65(6):275-278
4. Bonebrake RG, Watson P, Lanspa SJ. Upper gastrointestinal bleeding in acute pancreatitis. J Clin Gastroenterol. 1992;14(3):274-275
5. UK Working Party on Acute Pancreatitis. UK guidelines for the management of acute pancreatitis. Gut. 2005 ;54 Suppl 3: 1-9.
6. Bradley EL III. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis; September 11-13, 1992; Atlanta, GA. Arch Surg. 1993;128:586-590
7. Whitcomb DC. Clinical practice. Acute pancreatitis. N Engl J Med 2006; 354: 2142-2150
8. Pezzilli R, Zerbi A, Di Carlo V, Bassi C, Delle Fave G.F. Practical Guidelines for Acute Pancreatitis. Pancreatology 2010;10(5):523-535
9. Bank S, Indaram A. Causes of acute and recurrent pancreatitis. cClinical
14
Accepted Article
considerations and clues to diagnosis. Gastroenterol Clin North Am 1999; 28: 571-589
10. Balsarkar DJ, Joshi MA. Rupture of splenic artery pseudoaneurysm presenting with massive upper gastrointestinal bleed. Am J Surgery 2002;183:197–198
11. Lee KM, Paik CN, Chung WC, Yang JM. Association between acute pancreatitis and peptic ulcer disease. World J Gastroenterol 2011;17(8):1058-1062
12. Bank S, Pooran N, Singh R.The Prognosis of Gastrointestinal Hemorrhage in Severe Acute Pancreatitis has Improved Over the Last 20' Years. AGA;S1786
13. Rahman SH, Ammori BJ, Holmfield J, Larvin M, McMahon MJ. Intestinal hypoperfusion contributes to gut barrier failure in severe acute pancreatitis. J Gastrointest Surg 2003;7(1):26-36
14. Besselink MG, van Santvoort HC, Boermeester MA, Fischer K, Renooij W, de Smet MB, Cirkel GA, van Ramshorst B, Weusten BL, Schaapherder AF, Witteman BJ, Ploeg RJ, van Goor H, van Laarhoven CJ, Rosman C, Brink MA, van der Harst E, Wahab PJ, van Eijck CH, Dejong CH, Van Erpecum KJ, Akkermans LM, Gooszen HG. Gastrointestinal permeability and infectious complications in acute pancreatitis; a prospective multicenter study. AGA Abstracts: A-119
15. Cosen-Binker LI, Binker MG, Negri G, Tiscornia O. Influence of stress in acute pancreatitis and correlation with stress-induced gastric ulcer Pancreatology 2004;4(5):470-484
16. Elton Cahow C, Gusberg RJ, Gottlieb LJ. Gastrointestinal hemorrhage from pseudoaneurysms in pancreatic pseudocysts. Am J Surg 1983;145(4):534-541
15
Accepted Article
17. Kubo K, Nakamura H, Hirohata Y, Abe S, Onari N, Otsuki M. Ruptured aneurysm and gastric perforation associated with acute pancreatitis: a rare cause of hematemesis. Gastrointest Endosc 2001;53:658–60.
18. Balsarkar DJ, Joshi MA. Rupture of splenic artery pseudoaneurysm presenting with massive upper gastrointestinal bleed. Am J Surg 2002;183:197–198
19. Bedioui H, Ayadi S, Daghfous A, Jouini M, Bakhtri M, Rajhi H, Chebbi Faouzi, Fteriche F, Ksantini R, Kacem MJ, Safta ZB, Pseudoaneurysm of the splenic artery presenting with gastrointestinal bleeding. J Emerg Med 2010;38(3):317–319
20. Dasgupta R, Davies NJ, Williamson RCN, Jackson JE. Haemosuccus panreaticus: treatment by arterial embolizasion. Clin Radiol 2002;57:1021-1027
21. Hagiwara A, Miyauchi H, Shimazaki S. Predictors of vascular and gastrointestinal complications in severe acute pancreatitis. Pancreatology 2008;8(2):211-218
22. Hsu CY, Lee KC, Chan CC, Lee FY, Lin HC. Gastric necrosis and perforation as a severe
complication
of
pancreatic
pseudocyst
J
Chin
Med
Assoc
2009;72(11):603–606
23. Alhazzani W, Alshahrani M, Moayyedi P, Jaeschke R. Stress ulcer prophylaxis in critically ill patients: review of the evidence. Pol Arch Med Wewn 2012;122 (3): 107-114
24. Marik PE, Vasu T, Hirani A, Pachinburavan M. Stress ulcer prophylaxis in the new millennium:
a
systematic
review
and
meta-analysis.
Crit
Care
Med
2010 ;38(11):2222-2228
Table 1. Comparison of the patient characteristics and etiologies of Severe Acute Pancreatitis
16
Accepted Article
(SAP) with Upper Gastrointestinal Bleeding (UGIB) and those without UGIB (Non-UGIB) UGIB group
Non-UGIB group
(n=18)
(n=83)
55.78±14.78
51.88±16.82
0.3655
Men
14(77.8%)
51(61.4%)
0.2784
Women
4(22.2%)
32(38.6%)
25.22± 5.20
24.45±4.75
P-value
Patient characteristics Age (yrs) Gender (male/female)
Body mass index (kg/m2)
0.5581
Etiologies* Biliary
0.5902
No
8(44.4%)
29(34.9%)
Yes
10(55.6%)
54(65.1%)
Alcoholic
0.5216
No
13(72.2%)
67(80.7%)
Yes
5(27.8%)
16(19.3%)
Hypertriglyceridemia
0.7111
No
15(83.3%)
72(86.7%)
Yes
3(16.7%)
11(13.3%)
Other etiologies
0.1493
No
15(83.3%)
78(94.0%)
Yes
3(16.7%)
5(6.0%)
8(44.4%)
9(10.8%)
Mortality
0.0021
Note: *The reason the total number of different etiology groups was higher than the sum of the number in either the UGIB or non-UGIB groups was that several patients had more than one etiology
Table 2. Comparison of patients with Upper Gastrointestinal Bleeding (UGIB) and those without 17
Accepted Article
UGIB (Non-UGIB) UGIB group
Non-UGIB group
(n=18)
(n=83)
APACHE II score
17.61±6.02
9.35±6.67
Prevalence and risk factors for clinically significant upper gastrointestinal bleeding in patients with severe acute pancreatitis1
Xian-Bao Zhan, Xiao-Rong Guo ,Jing Yang, Jie Li, Zhao-Shen Li
Department of Gastroenterology, Changhai Hospital, Second Military Medical University, Shanghai 200433, China
Correspondence to: Dr. Zhaoshen Li, Professor of Medicine, Chief of Department of Gastroenterology, Changhai Hospital, Second Military Medical University, Shanghai, 200433, China E-mail: [email protected]
Financial disclosure and conflict of interest: None of the authors have commercial relationships that might pose a conflict of interest in connection with this manuscript.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between
this version and the Version of Record. Please cite this article as doi: 10.1111/1751-2980.12206 1
Accepted Article
Abstract
Goals: To explore the prevalence and risk factors of upper gastrointestinal bleeding(UGIB) in patients with severe acute pancreatitis(SAP). Background: The treatment strategy of SAP changed greatly during the past decade,
but the prevalence and risk factors of the complicated UGIB are still unclear. Study: From July 2006 to June 2010, retrospective data on 101 patients with SAP admitted to Changhai Hospital were collected. The prevalence and risk factors of UGIB were analyzed. Results: Eighteen patients(17.8%) developed UGIB. Thirteen patients received
endosopic examinations which yielded 6 cases of acute gastric mucosal lesions(AGML), and 5 cases of peptic ulcer. The mortality rate of UGIB patients was much higher than that in non-UGIB patients (44.4% vs 10.8%, P=0.0021). Univariate analysis revealed that the risk factors for UGIB included acute physiology and chronic health evaluation II (APACHE II) score, computed tomography severity index, Ranson score, arterial blood pH values and oxygen pressures (PaO2), serum blood urea nitrogen
and
creatinine
concentrations,
prothrombin
and
activated
partial
thromboplastin times, platelet counts, the presence of shock, sepsis and organ failure other than UGIB, mechanical ventilation, heparinized continuous renal replacement therapy and total parental nutrition. By multivariate logistic regression analysis, APACHE II scores and CT severity indices(CTSI) were found to be significant risk factors, while PaO2 was the only protective factor. Conclusions: UGIB is a common complication with a poor prognosis due mainly to
2
Accepted Article
peptic ulcer and AGML. Patients with high APACHE II scores, and CTSI or low PaO2 should be considered to be at high risk for UGIB.
Key words: acute pancreatitis;upper gastrointestinal bleeding;risk factor;peptic ulcer;acute gastric mucosal lesion
3
Accepted Article
1. Introduction 2. Severe acute pancreatitis (SAP) occurs in approximately 20% of cases of acute pancreatitis, and is associated with a 15% to 30% mortality rate [1]. The outcome of SAP mostly depends on whether systemic and local complications develop. It has been recognized that upper gastrointestinal bleeding (UGIB) is one of the most common complications with SAP [2-4]. Previous reports have shown that 50–65% of patients with acute pancreatitis may have acute gastrointestinal mucosal lesions [2-3]. According to the definition established at the Atlanta Symposium in 1992 and UK guidelines for the management of acute pancreatitis in 2005, gastrointestinal bleeding greater than 500 mL in 24 h can be diagnosed as gastrointestinal failure [5,6]. During the last decade, the treatment of SAP has
changed greatly. Currently, conservative therapy is the treatment of choice for SAP [7-8]. Although the prophylactic use of proton pump inhibitors been widely adopted in SAP treatment regimens, the current prevalence and risk factors of UGIB in patients with SAP treated prophylactically with proton pump inhibitors is
unclear. This study aimed to shed light on this issue.
4
Accepted Article
2. Materials and Methods
We conducted a retrospective chart review of patients admitted to the specialized pancreatic intensive care unit at Changhai Hospital between July 2006 and June 2010. Changhai Hospital is a tertiary and academic hospital, affiliated to Second Military
Medical University, and serves patients from various parts of China.
2.1 Data Collection
Charts were studied and data was retrospectively collected using a standardized data collection form. Collected variables included patient demographics, medical history, etiology of acute pancreatitis, laboratory tests (white blood cell count, red blood cell count, platelet count, hemoglobin, serum liver tests, serum creatinine and blood urea
nitrogen, prothrombin time, arterial blood gas analysis, etc.), systemic and local complications (systemic inflammatory response syndrome (SIRS), organ failure, acute fluid collection, pseudocyst, pancreatic abscess), special treatment (continuous blood purification, peritoneal lavage, mechanical ventilation, total parental nutrition (TPN), steroids use, etc.) and primary outcomes (transfer to surgery, recovery or death). The
definitions of systemic and local complications were based on the clinical classification system for acute pancreatitis established at the Altanta Symposium in 1992 and UK guidelines for the management of acute pancreatitis in 2005 [5,6].
5
Accepted Article
2.2 Patient Selection
The enrolled SAP patients were all admitted to the specialized pancreatic intensive care unit at Changhai Hospital from July 2006 to June 2010. The diagnosis of acute pancreatitis was made on clinical criteria which included elevated serum amylase (>3 times of upper limit of normal), and evidence of pancreatitis on ultrasonography or computed tomography (CT). SAP was defined as an (APACHE II) of 8 or more, or Ranson score of 3 or more [1, 6-7]. All patients with SAP underwent detailed clinical
evaluation and investigation, including monitoring of vital signs, hematology and serum biochemistry tests, arterial blood gas analysis, and plain film radiography of the abdomen and chest. A dynamic contrast-enhanced CT was done around 72 h after the onset of SAP. During the first 72 h, active fluid resuscitation with colloids was performed. During the acute phase, usually within five to seven days after onset, proton pump inhibitors, somatostatin, gabexate and wide spectrum antibiotics were routinely used. ERCP and nasal-biliary drainage were conducted within the 72 h if acute biliary pancreatitis was diagnosed or if biliary infection existed any time. If the systemic inflammatory response syndrome existed even after conservative therapy, or if the diagnosis of hypertriglyceridemia was made, continuous renal replacement therapy (CRRT) was performed. Clinically-significant bleeding (CSB) was defined by the presence of coffee
ground gastric aspirate, gross hematemesis, or melena, with or without circulatory effects.
The
etiologies
were
classified
6
into
four
major
groups,
biliary,
Accepted Article
hypertriglyceridemia, alcoholic and other etiologies [9]. Acute biliary pancreatitis was
defined as serum alanine aminotransferase elevations >3 times of the upper limit of the normal value with or without elevations of liver function tests. Acute alcoholic pancreatitis was defined as acute pancreatitis that developed in the context of an oral intake of more than 20 g of alcohol during the past 24 hours. Acute hypertriglyceridemia-related pancreatitis was defined as acute pancreatitis that developed in the presence of a plasma triglyceride concentration over 5.65 mmol/L.
2.3. Statistical Analysis
All analyses were performed using SAS9.1 software. Frequencies, percentages, means, standard deviations, odds ratios, and confidence intervals were calculated. Continuous data were expressed as mean ± SE (standard error of the mean) determined using the Student’s t-test, while categorical variables were expressed as quantities, and were analyzed using the χ2 test. Multiple stepwise logistic regression analysis was used to identify the risk factors for UGIB. Differences in means were conducted using the
2-tailed Student’s t-test, and differences in percentages were assessed using the χ2 test. Continuous variables were tested using logistic regression methods. P values of 0.05). The median interval between
the onset of SAP and UGIB was 6 days. The mortality rate of the UGIB group was much higher than that of the non-UGIB group (P=0.0021) (Table 1).
3.2 Causes of UGIB Of the 18 patients in the UGIB group, 13 cases received upper gastrointestinal
endoscopic examinations during the hospital stay period which revealed 2 cases of gastric ulcer, 6 cases of acute gastric mucosal lesions, 3 cases of duodenal ulcer, and 2 cases of pancreatic necrotic tissue invading into the duodenal bulb presenting as a multi-lesion, honeycomb-like ulcer (Figure 1).
3.3 Risk Factors of UGIB in Patients with SAP On univariate analysis, APACHE II score, CT severity index, Ranson score, arterial blood pH value and PaO2, serum blood urea nitrogen and creatinine, prothrombin time and activated partial thromboplastin time, platelet counting, shock, sepsis, organ failure other than UGIB, mechanical ventilation time, duration of heparinized CRRT, duration of TPN, were found to be significantly associated with UGIB in patients with acute
8
Accepted Article
pancreatitis (P0.05) (Table 2). On multivariate logistic regression analysis, only the APACHE II score and CT
severity index were found to be significant risk factors, while arterial blood PaO2 was
the only protective factor (Table 3).
9
Accepted Article
4. Discussion SAP can deteriorate very rapidly. UGIB, one of the common complications of SAP, may be lethal and complicate the course of SAP. In the current study, the mortality rate of UGIB patients was found to be much higher than that in non-UGIB patients (44.4% vs. 10.8%, P=0.0021). The exact incidence of UGIB in SAP is currently unknown.
Sharma et al. reported that gastrointestinal hemorrhage occurred in 3.6% of patients
with acute pancreatitis [5]. In the current study, we found that UGIB occurred in 17.82% of patients with SAP, which is much higher than that reported by Sharma et al. The reason for the difference in incidence is likely due to that fact that the severity of
disease of our patients was greater than that reported by Sharma et al. The APACHE II score in our group was 17.61±6.02 in the UGIB group, and 9.35±6.67 in the non-UGIB group, while it was 8.7±3.2 in the gastrointestinal hemorrhage group and 5.3±5.2 in the
non-gastrointestinal hemorrhage group reported by Sharma et al. In patients with SAP, a variety of causes of UGIB have been reported including
acute gastric mucosal lesions and stress ulcer, peptic ulcer, gastric variceal bleeding,
pseudoaneurysm rupture into the duodenum or into the pancreatic duct, nasogastric tube damage, etc. [11-12]. Acute gastric mucosal lesions in patients with SAP can range from simple gastritis and erosions to ulceration or bleeding. It has been reported that the majority of the lesions or ulcers were located in the stomach and duodenum, while lesions in the esophagus were uncommon [3]. In our study, thirteen patients with SAP and UGIB received endoscopic examinations which showed 6 cases of acute
gastric mucosal lesions and 5 cases of gastric ulcer and duodenal ulcer. The results
10
Accepted Article
were consistent with those of a previous report [3]. Stress of severe illness, hypoxia,
ventilation and non-steroidal anti-inflammatory drugs for pain relief, and gastrointestinal hypoperfusion have been proposed to be potential causes of gastrointestinal mucosal barrier failure, and to contribute to the development of acute
mucosal lesions or ulcers [11, 13-14]. Animal studies have shown that prior restraint
stress aggravated pancreatic pathology from biliary-pancreatic duct outlet exclusion in closed duodenal loop models. Acute pancreatitis has been shown to exacerbate gastric damage caused by restraint stress resulting in ulceration evolving to ulcer, hemorrhage and gastric necrosis [15]. Gastrointestinal hemorrhages secondary to hemosuccus
pancreaticus or ruptured artery pseudoaneurysm are rare conditions that pose significant diagnostic and therapeutic challenges. There are a number of case reports related to such conditions [16-20]. They have been reported to occur most commonly in a setting of acute pancreatitis with rupture of pseudoaneurysms of the splenic or hepatic artery into the pancreatic duct, which were definitively diagnosed by selective celiac angiography [16]. Severe gastroduodenal necrosis with perforation has rarely
been reported [21-22]. Kubo et al. [19] described a patient with acute pancreatitis who presented with hematemesis and gastric perforation caused by gastric erosion and rupture of a pseudoaneurysm. In the current study, we found 2 cases of duodenal necrosis with perforation with refractory bleeding which resulted from erosion of pancreatic necrosis into the duodenal wall. It is difficult for clinicians to recognize this condition. A definitive diagnosis can be established by endoscopic examination. There are many vascular structures neighboring the duodenal segment, and erosion of these
11
Accepted Article
vessels may result in fatal hemorrhage. Because of the severe consequences,
physicians should be alert and consider it in the differential diagnosis. The risk factors for hemorrhagic complications of UGIB in critically ill patients
have been previously studied [23-24]. It has been proposed that old age, trauma, surgery, shock, infection, sepsis, severe jaundice, multiorgan dysfunction syndrome (MODS) or multiple systemic organ failure (MSOF), coagulation disorders, a history of ulcers during the year, long-term fasting and nutritional deficiencies were risk factors for UGIB. Although SAP is also a critical illness, there are very few reports on the risk factors for UGIB in patients with SAP within the last two decades. In the report by Sharma et al., necrotizing pancreatitis, high APACHE II score, organ failure, sepsis and pancreatic collection were significant risk factors for developing hemorrhage. However, that study included both gastrointestinal and abdominal
hemorrhages. Hagiwara et al. performed a prospective cohort study on patients with SAP to determine prognostic factors for arterial injury and gastrointestinal perforation in patients with SAP, and found that CRP ≥15 mg/dl on day 7, and a CT severity index ≥7 between days 6 and 8 were independent risk factors (P = 0.02 and 0.04, respectively). The odds ratios for CRP ≥15 mg/dl were 23.0 and 15.7, respectively for a CT severity index of ≥7. The results suggest that a persistent elevation of the CRP concentration and a high CT severity index are independent risk factors for local complications associated with SAP [16]. The current study was the first specifially designed to explore the risk factors of UGIB in patients with SAP. In the current study, we found the occurrence of UGIB had no direct relationship to the etiology of SAP,
12
Accepted Article
which was consistent with a previous report [5]. The conditions related to the risk factors in the current study actually reflect the
severity of the disease, which may have induced the gastroduodenal mucosal lesions by destroying the mucosal barrier or systemic coagulopathy. This hypothesis needs to be confirmed by further basic research. Unexpectedly, the current study found that the
liver function tests and age were not statistically different between the UGIB group and the non-UGIB group, which is not consistent with the findings of previous studies on the risk factors of stress ulcer bleeding in critically ill patients which included a variety of diseases. The reason for this needs further study, but one possible explanation might
be that the transient abnormal liver function indices do not reflect the degree of liver function loss under the conditions of acute biliary pancreatitis in the current study. The potential limitations of the current study are that it involved only a single
center, and was retrospective with a small number of patients. However, all the enrolled patients in this study were treated by the same medical group, which helped ensure that all the enrolled patients underwent similar basic management, and also allowed control of the study quality.
In conclusion, UGIB is an early common complication in patients with severe acute pancreatitis and is associated with a high mortality rate. The patients with a high APACHE II score and a CT severity index or low PaO2 should be considered at high risk for UGIB.
13
Accepted Article
References 1. Banks PA, Freeman ML. Practice guidelines in acute pancreatitis. Am J Gastroenterol 2006; 101: 2379-2400
2. Chen TA, Lo GH, Lin CK, Lai KH, Wong HY, Yu HC, Hsu PI, Chen HH, Tsai WL, Chen WC. Acute pancreatitis-associated acute gastrointestinal mucosal lesions: incidence, characteristics, and clinical significance. J Clin Gastroenterol 2007; 41:630-634
3. Lin CK, Wang ZS, Lai KH, Lo GH, Hsu PI. Gastrointestinal mucosal lesions in patients with acute pancreatitis. Zhonghua Yi Xue Za Zhi (Taipei). 2002;65(6):275-278
4. Bonebrake RG, Watson P, Lanspa SJ. Upper gastrointestinal bleeding in acute pancreatitis. J Clin Gastroenterol. 1992;14(3):274-275
5. UK Working Party on Acute Pancreatitis. UK guidelines for the management of acute pancreatitis. Gut. 2005 ;54 Suppl 3: 1-9.
6. Bradley EL III. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis; September 11-13, 1992; Atlanta, GA. Arch Surg. 1993;128:586-590
7. Whitcomb DC. Clinical practice. Acute pancreatitis. N Engl J Med 2006; 354: 2142-2150
8. Pezzilli R, Zerbi A, Di Carlo V, Bassi C, Delle Fave G.F. Practical Guidelines for Acute Pancreatitis. Pancreatology 2010;10(5):523-535
9. Bank S, Indaram A. Causes of acute and recurrent pancreatitis. cClinical
14
Accepted Article
considerations and clues to diagnosis. Gastroenterol Clin North Am 1999; 28: 571-589
10. Balsarkar DJ, Joshi MA. Rupture of splenic artery pseudoaneurysm presenting with massive upper gastrointestinal bleed. Am J Surgery 2002;183:197–198
11. Lee KM, Paik CN, Chung WC, Yang JM. Association between acute pancreatitis and peptic ulcer disease. World J Gastroenterol 2011;17(8):1058-1062
12. Bank S, Pooran N, Singh R.The Prognosis of Gastrointestinal Hemorrhage in Severe Acute Pancreatitis has Improved Over the Last 20' Years. AGA;S1786
13. Rahman SH, Ammori BJ, Holmfield J, Larvin M, McMahon MJ. Intestinal hypoperfusion contributes to gut barrier failure in severe acute pancreatitis. J Gastrointest Surg 2003;7(1):26-36
14. Besselink MG, van Santvoort HC, Boermeester MA, Fischer K, Renooij W, de Smet MB, Cirkel GA, van Ramshorst B, Weusten BL, Schaapherder AF, Witteman BJ, Ploeg RJ, van Goor H, van Laarhoven CJ, Rosman C, Brink MA, van der Harst E, Wahab PJ, van Eijck CH, Dejong CH, Van Erpecum KJ, Akkermans LM, Gooszen HG. Gastrointestinal permeability and infectious complications in acute pancreatitis; a prospective multicenter study. AGA Abstracts: A-119
15. Cosen-Binker LI, Binker MG, Negri G, Tiscornia O. Influence of stress in acute pancreatitis and correlation with stress-induced gastric ulcer Pancreatology 2004;4(5):470-484
16. Elton Cahow C, Gusberg RJ, Gottlieb LJ. Gastrointestinal hemorrhage from pseudoaneurysms in pancreatic pseudocysts. Am J Surg 1983;145(4):534-541
15
Accepted Article
17. Kubo K, Nakamura H, Hirohata Y, Abe S, Onari N, Otsuki M. Ruptured aneurysm and gastric perforation associated with acute pancreatitis: a rare cause of hematemesis. Gastrointest Endosc 2001;53:658–60.
18. Balsarkar DJ, Joshi MA. Rupture of splenic artery pseudoaneurysm presenting with massive upper gastrointestinal bleed. Am J Surg 2002;183:197–198
19. Bedioui H, Ayadi S, Daghfous A, Jouini M, Bakhtri M, Rajhi H, Chebbi Faouzi, Fteriche F, Ksantini R, Kacem MJ, Safta ZB, Pseudoaneurysm of the splenic artery presenting with gastrointestinal bleeding. J Emerg Med 2010;38(3):317–319
20. Dasgupta R, Davies NJ, Williamson RCN, Jackson JE. Haemosuccus panreaticus: treatment by arterial embolizasion. Clin Radiol 2002;57:1021-1027
21. Hagiwara A, Miyauchi H, Shimazaki S. Predictors of vascular and gastrointestinal complications in severe acute pancreatitis. Pancreatology 2008;8(2):211-218
22. Hsu CY, Lee KC, Chan CC, Lee FY, Lin HC. Gastric necrosis and perforation as a severe
complication
of
pancreatic
pseudocyst
J
Chin
Med
Assoc
2009;72(11):603–606
23. Alhazzani W, Alshahrani M, Moayyedi P, Jaeschke R. Stress ulcer prophylaxis in critically ill patients: review of the evidence. Pol Arch Med Wewn 2012;122 (3): 107-114
24. Marik PE, Vasu T, Hirani A, Pachinburavan M. Stress ulcer prophylaxis in the new millennium:
a
systematic
review
and
meta-analysis.
Crit
Care
Med
2010 ;38(11):2222-2228
Table 1. Comparison of the patient characteristics and etiologies of Severe Acute Pancreatitis
16
Accepted Article
(SAP) with Upper Gastrointestinal Bleeding (UGIB) and those without UGIB (Non-UGIB) UGIB group
Non-UGIB group
(n=18)
(n=83)
55.78±14.78
51.88±16.82
0.3655
Men
14(77.8%)
51(61.4%)
0.2784
Women
4(22.2%)
32(38.6%)
25.22± 5.20
24.45±4.75
P-value
Patient characteristics Age (yrs) Gender (male/female)
Body mass index (kg/m2)
0.5581
Etiologies* Biliary
0.5902
No
8(44.4%)
29(34.9%)
Yes
10(55.6%)
54(65.1%)
Alcoholic
0.5216
No
13(72.2%)
67(80.7%)
Yes
5(27.8%)
16(19.3%)
Hypertriglyceridemia
0.7111
No
15(83.3%)
72(86.7%)
Yes
3(16.7%)
11(13.3%)
Other etiologies
0.1493
No
15(83.3%)
78(94.0%)
Yes
3(16.7%)
5(6.0%)
8(44.4%)
9(10.8%)
Mortality
0.0021
Note: *The reason the total number of different etiology groups was higher than the sum of the number in either the UGIB or non-UGIB groups was that several patients had more than one etiology
Table 2. Comparison of patients with Upper Gastrointestinal Bleeding (UGIB) and those without 17
Accepted Article
UGIB (Non-UGIB) UGIB group
Non-UGIB group
(n=18)
(n=83)
APACHE II score
17.61±6.02
9.35±6.67
