by Harwood Academic Publishers Printed in Singapore Endoscopic Nasobiliary Drainage in the Management of Acute Cholangitis: An Experience in 143 Patients

Acute cholangitis is associated with a high mortality and morbidity and often requires drainage of the obstructed biliary system. The purpose of this study was to evaluate the usefulness and safety of endoscopic nasobiliary drainage in the treatment and prevention of acute cholangitis due to diverse etiology. During a 32-month period, 143 patients (67 males, 76 females) with age range of 15 to 84 years underwent urgent fluoroscopy guided endoscopic nasobiliary drainage using a 7 Fr catheter either to treat acute cholangitis not responding to antibiotics (group A, n = 116) or to prevent its development following endoscopic retrograde cholangiography performed in an obstructed biliary system (group B, n = 27). Underlying etiology included bile duct stones (92), malignant biliary obstruction (34), choledochal cyst (4), chronic pancreatitis (4), ruptured hydatid cyst (3), portal hypertensive cholangiopathy (3) and liver abscess (3). Endoscopic nasobiliary drainage was performed successfully in 129 patients (90.2%). Cholangitis improved within 1 to 3 days (in group A) or did not develop (in Group B) in 125 patients (96.7%) with successful endoscopic nasobiliary drainage. Two patients however required additional drainage by percutaneous transhepatic route, while two died inspite of effective endoscopic drainage. Of the 14 patients (9.8%) with failed endoscopic drainage, 9 were managed by surgical decompression or percutaneous transhepatic drainage, 3 died of septicemia. Endoscopic nasobiliary drainage is a safe and effective method to treat patients with acute cholangitis as well as to prevent its development following cholangiography performed in an obstructed biliary system.


MATERIAL AND METHODS
Between May, 1993 andDecember, 1995, 143 patients underwent endoscopic nasobiliary drainage (ENBD) either to treat acute cholangitis (Group A, therapeutic group, n 116) or to prevent the risk of cholangitis following endoscopic retrograde cholangiopancreatography (ERCP) performed in an obstructed biliary system with failure to relieve the obstruction endoscopically (Group B, prophylactic group, n 27).
All patients underwent clinical evaluation, laboratory investigations including liver function tests, renal parameters and serum electrolytes. An abdominal ultrasonogram was also performed in all patients. Initial treatment of patients with features of cholangitis (Group A) included intravenous antibiotic (ciprofloxacin or piperacillin), parenteral fluid and injections of vitamin K, while those without cholangitis (Group B) were started on oral ciprofloxacin 12-24 hours prior to ERCP. Four patients in Group A having hyperkalemia due to renal failure secondary to cholangitis and septicemia received dialysis prior to ERCP. ERCP was performed in Group A patients only after a sub-optimum response to medical therapy instituted for a period of 1 to 3 days.
ERCP was performed under fluoroscopic guidance using a side-viewing duodenoscope (JF-IT or JF-IT20, Olympus or FD34X, Ashai Optical) after intravenous diazepam (5-10 mg) and/or hyoscine N-butyl bromide (20-60 mg). A small amount of 60% meglumine iothalamate was used for diagnostic ERCP. Endoscopic sphincterotomy was performed in only 11 patients as a pre-requisite for attempted stone extraction. A 0.035 inch guidewire (Zebra, Microvasive, Boston Scientific, USA) was passed through the ERCP cannula and after positioning its tip proximal to the site of obstruction, the cannula was withdrawn. In patients with tight biliary stricture, a 6 Fr biliary dilator (Wilson Cook, Winston-Salem) was passed over the guide wire. Nasobiliary catheter (7 Fr, pig-tail) was threaded over the guide wire, passed through the biopsy channel of the endoscope and its tip positioned proximal to the site of obstruction or into the abscess cavity (in patients with liver abscess, hydatid cyst and choledochal cyst). Guide wire was then withdrawn followed by the endoscope. After ensuring a free flow of bile from the external end of ENBD catheter, the catheter was rerouted through the nose by a rail-road technique using a nasogastric tube. A cholangiogram was repeated through the ENBD catheter using 60% meglumine iothalamate to adequately visualise the bitiary system. ENBD catheter, while in position, was irrigated once daily using 20 ml-60 ml of sterile normal saline.
Response to endoscopic drainage was monitored in terms of clinical improvement as well as changes in liver function tests. Patients were offered definitive therapy for the underlying etiology, once the cholangitis settled and patient stabilised. *Includes gall bladder cancer with bile duct infiltration as well as cholangiocarcinoma **Two of these patients ahd cholangitis following occlusion of biliary stent ***Failed stone extraction because of large stones (4), associated stricture (2), abnormal coagulogram (1)

RESULTS
There were 67 men and 76 women with an age range of 15 to 84 years. Table I gives the underlying etiology of cholangitis or biliary obstruction needing ENBD.
Among patients with cholangitis, common bile duct stones constituted the major proportion (n 83, 71.6%) ( Figure 7 shows the outcome of ENBD placement. Nasobiliary catheter could be positioned successfully in 129 patients (90.2%). Failure in 14 patients (malignant obstruction: 9, chronic pancreatitis: 1, stone disease: 4) was due to failure to cannulate the bile duct in 3 patients and failure to negotiate guide wire across the obstruction in 11 patients due to tight stricture (n 8), multiple stones (n 2) or a large stone (n 1). Of the 8 patients with tight biliary stricture and failure to negotiate the guide wire, one had benign stricture due to chronic pancreatitis, two had ampullary carcinoma and five had high bile duct malignant stricture either due to infiltration from carcinoma of gall bladder or due to cholangiocarcinoma. While 5 of these patients with failed ENBD underwent percutaneous transhepatic biliary drainage, 4 underwent biliary tract surgery. Among the failed ENBD group, 3 patients died due to uncontrolled cholangitis and septicemia, 1 of these inspite of percutaneous drainage.
Of the 129 patients with successful placement of ENBD; 125 (96.9%) had control of cholangitis within 1 to 3 days (Group A) or had successful prevention of cholangitis (Group B). Six of these patients did require repositioning of the catheter due to its being pulled out by the patient (n 3) or getting kinked in the stomach (n 3). Two patients had inadequate response and required additional percutaneous transhepatic biliary drainage. Remaining two patients died inspite of (A) (B) successful ENBD; one of these had a choledochal cyst and patient succumbed to septicemic shock within minutes of ENBD; the other patient had ENBD placed for malignant biliary stricture but had pulled out her catheter and died of septicemia before the catheter could be repositioned. In patients with favourable response to ENBD, catheter was removed after a period of 5 to 14 days (median: 8 days).
Of the surviving 127 patients with ENBD, 76 were given a more definitive treatment for the underlying biliary disease after stabilisation of patient's status. This included bile duct stone extraction after endoscopic sphincterotomy in 40 patients, biliary surgery in 28 and endoscopic biliary stent placement in 8.

DISCUSSION
Endotherapy in acute cholangitis can be in the form of endoscopic sphincterotomy, biliary stenting or by nasobiliary drainage (ENBD). In the present series, we could establish ENBD in 90.2% of the patients and successful ENBD led to a rapid control of cholangitis or prevention of cholangitis in 97% of patients. The procedure was safe with no morbidity or mortality related to ENBD. The overall short-term mortality in our patient group was 3.5 %, being 1.5% only in those with successful ENBD. Death was thus, always related to delay in institution of ENBD, displacement of catheter or failure to establish ENBD. Our results with ENBD are similar to the earlier reports with this procedure [8][9][10][11][12][13]. A (M (B) FIGURE 3 Patient with porta hepatitis block due to malignancy (a) Following ERCP, nasobiliary drain was established without sphincterotomy to prevent cholangitits (b) Subsequently sphincterotomy was performed followed by dilatation of porta block by biliary dilator (arrows) (c) A 10 Fr stent was placed.
slightly higher failure rate in the present study could be due to the fact that we have included patients with diverse etiology including malignant bile duct obstruction. ENBD failure were mostly in malignant etiology group because of tightness of these strictures, resulting in difficulty to negotiate the guide wire across the stricture. Among the patients with malignant bile duct strictures, those having high bile duct strictures resulting from cholangiocarcinoma or due to infiltration from carcinoma of gall bladder had a relatively higher failure rate for guide wire negotiation and ENBD placement, an experience similar to the earlier reports [15]. While as expected choledocholithiasis was the commonest benign cause for cholangitis, we unlike earlier series [9][10][11][12][13][14] also encountered other causes such as chronic pancreatitis with biliary stricture, choledochal cyst, hydatid cyst with biliary rupture and portal cholangiopathy. ENBD resulted in control of cholangitis and stabilised the patient, (C) so that a more definite therapy in the form of surgery Endoscopic sphincterotomy followed by dormia extraction in the same session is often considered the best therapy for cholangitis due to bile duct stones since it combines drainage of infected system along with removal of the cause [8][9][10]. However, endoscopic sphincterotomy in acute cholangitis has been associated in earlier reported series with complications such as hemorrhage, perforation and pancreatitis in 6-12% patients and mortality in 4.7-7.6% [8,10]. Moreover sphincterotomy may not always be possible in patients with cholangitis due to abnormal coagulogram, sphincterotomy combined with stone extraction by basket or balloon is more time consuming than ENBD placement without sphincterotomy [7,9,14] and ductal clearance after sphincterotomy has been achieved in earlier reports in 66 to 76% patients only [8][9][10]. In the present series, ENBD was performed without any prior sphincterotomy in 91.5% patients.
Some workers have preferred biliary stenting over ENBD while managing acute cholangitis because of patient's convenience with the former [ 16]. ENBD was preferred by us because the indwelling catheter permitted a cholangiogram to be performed later, allowing us to inject minimal required contrast material during initial procedure in the presence of infected bile. This reduces the risk of increased cholangiovenous reflux which can aggravate or precipitate septicemia in a patient with cholangitis [1,9]. ENBD moreover, allowed collection of bile for culture and flushing of catheter for clearing flakes of pus and bile debris. ENBD catheter can also act as a conduit for chemical dissolution of gall stones [I,14]. Though ENBD does carry the risk of being pulled out by the patient as happened in 3 of our cases, it can almost always be repositioned rapidly. Combining ENBD with stent placement in patients having cholangitis due to malignant biliary obstruction can cure cholangitis and at the same time provide the palliation for the tumor obviating the need for subsequent endoscopic (A) (B) procedure. However, placement of two prosthetic device would require endoscopic sphincterotomy and a longer procedure time. We have preferred to treat the cholangitis in such setting by ENBD and a more definitive therapy by stenting or surgery was offered subsequently. Further, studies are however warranted to choose the optimum management modality in such a setting.
Percutaneous transhepatic biliary drainage (PTBD) and surgical treatment have been used earlier for treating patients with acute cholangitis. However, direct surgical intervention in acute cholangitis has been reported to carry a mortality of 6.5-40% [2][3][4]. PTBD is similarly associated with a significant mortality and a high risk (upto 28%) of bleeding, bile leakage and peritonitis and occasionally of pneumothorax, traumatic pseudoaneurysm and arterioportal fistula formation [5][6][7]17]. We therefore feel that surgery or PTBD should be performed only when ENBD is not possible due to anatomical reasons such as previous Polya's gastrectomy, is technically unsuccessful or (C) rarely when ENBD does not improve the cholangitis. PTBD was performed in 7 patients in the present series, in 5 patients beuse of failed ENBD and in 2 with successful ENBD for more effective drainage.
In conclusion, the present study performed over a large number of patients and in a wide etiological spectrum confirms the usefulness and safety of ENBD in the treatment and prevention of acute cholangitis.