- Email: [email protected]
Complications of endoscopic sphincterotomy
GASTROENTEROLOGY 1991:101:1066-1075
Complications of Endoscopic Sphincterotomy A Prospective Series With Emphasis on the Increased Risk Associated With Sphincter of Oddi Dysfunction and Nondilated Bile Ducts STUART
SHERMAN,
THOMAS
A. RUFFOLO,
ROBERT
H. HAWES,
and GLEN A. LEHMAN Division of Gastroenterology/Hepatology, Indiana University Hospital, Indianapolis, Indiana
Mostly retrospective series with limited use of sphincter of Oddi manometry have indicated that early complications are more common when endoscopic sphincterotomy is performed for sphincter of Oddi dysfunction than for common duct stones. The current study was undertaken to prospectively evaluate the frequency and type of complications of endoscopic sphincterotomy performed for sphincter of Oddi dysfunction compared with endoscopic sphincterotomy performed for other conditions. Four hundred twenty-three patients underwent sphincterotomy for sphincter of Oddi dysfunction (166), common duct stone(s) (163),tumor (60),and miscellaneous reasons (34). Patients were observed in the hospital for at least 24 hours after the procedure, and 36-day follow-up data were obtained. The overall complication rate was 6.9%, but complications were more frequent when sphincterotomy was performed for sphincter of Oddi dysfunction than for all other indications (10.6% vs. 4.3%; P = 0.009). Precut sphincterotomy was more frequently required in the sphincter of Oddi dysfunction group (21.1% vs. 11.7%, P = 0.009)but was no more likely to result in a complication (6.2%)than standard sphincterotomy. The risk of a complication was considerable for a small-diameter common bile duct (15 mm), particularly when sphincterotomy was performed for sphincter of Oddi dysfunction (37.5%). The overall 30-day mortality rate was l.7%, but the procedure-related mortality rate was believed to be 0.2%. Itis concluded that endoscopic sphincterotomy for sphincter of Oddi dysfunction is more hazardous than for other conditions, particularly when a small common bile duct is present.
ndoscopic sphincterotomy (ES) is gaining acceptance as an effective therapy for patients with sphincter of Oddi dysfunction (SOD) (l-4). However, reports have indicated that early complications are more common with ES performed for SOD than for Additionally, the morcommon duct stones (2,3,5-g). tality rate (two times higher for SOD) and late stenosis rate after ES (approximately four times higher for SOD) are substantially higher in patients with SOD (8,9). Most of these data come from retrospective series involving small numbers of patients with limited use of manometry to document SOD. The current study was undertaken to prospectively evaluate the frequency and type of complications of ES performed for SOD compared with ES performed for other indications.
E
Materials
and Methods
Between March 1989 and August 1990, endoscopic retrograde cholangiopancreatography (ERCPs) was performed 1204 times at Indiana University Hospital. The subjects of the study were the 423 (35.1%) patients who also underwent endoscopic biliary sphincterotomy. Four hundred ten (97%) were hospitalized and observed for complications for at least 24 hours after the procedure. Seven patients were transferred back to their referring hospital after ES, whereas the remaining 6 were carefully observed
Abbreviations used in thispaper: ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy; SOD, sphincter of Oddi dysfunction; SOM, sphincter of Oddi manometry. o 1991 by the American Gastroenterological Association 0016-5085/91/$3.00
October 1991
COMPLICATIONS
OF ENDOSCOPIC SPHINCTEROTOMY
1069
for at least 6 hours in an outpatient setting and contacted by telephone follow-up. Thirty-day follow-up data were obtained by chart review or telephone interview. Patients who underwent pancreatic sphincterotomy of the major or minor papilla are not included in this series. Patients underwent preendoscopic assessment including hepatic and pancreatic chemistries, complete blood count, and coagulation studies. The prothrombin time was within 3 seconds of the control value or corrected to that time with vitamin K and/or fresh frozen plasma before the procedure in all patients. Tests for bleeding times were not performed routinely; however, they were performed and the bleeding time was required to be normal if the patient had taken platelet-affecting drugs. A platelet count of >5O,OOO/~L was required before ES. Patients were asked to discontinue use of salicylates and other platelet-affecting drugs 7 days before [in nonemergent cases] the ES and for 14 days after the ES. All patients received prophylactic antibiotics (cefoperazone was generally used unless the patient was allergic to cephalosporins) before the procedure (and for at least 24 hours after the procedure, if an obstructed biliary tree was found at ERCP). Patients were administered pharyngeal local anesthesia and sedated with IV diazepam and Demerol (Meperidine; Wyeth Laboratories, Philadelphia, PA) as needed. Duodenal relaxation was obtained with glucagon. Generally, 30% meglumine diatrizoate was used for cholangiography, whereas a 60% concentration was used for pancreatography. Full pancreatography was performed in SOD patients, but, often, in biliary stone patients, the pancreatic duct was only (intentionally) partially filled. Endoscopic sphincterotomy was performed using Olympus duodenoscopes (Olympus Corporation of America, Lake Success, NY) and Erlangen pull-type sphincterotomes. Sphincterotomes with variable nose and cutting wire lengths, as well as those used with and without guidewires, were used interchangeably and tallied together. A controlled cut was made in stepwise fashion using short pulses of blended current for the lower half of the incision and alternating blended and pure coagulation current for the cephalad half. The length of the incision was somewhat tailored to the indication for the sphincterotomy and the size of the common duct stone(s) if present. The cut was extended through 75% of the longitudinal fold in all patients and through 85%-90% of the fold in most patients. The entire longitudinal fold was cut if this fold was short ( < 10 mm) or if biliary stones > 10 mm were present. Sphincterotomies were performed by three experienced endoscopists (S.S., R.H., and G.L.). When selective cannulation of the common bile duct with the standard sphincterotome was not achieved, a precut sphincterotomy was performed using a needle knife (lo), usually (in 63% of patients) over a previously placed 5-7F pancreatic stent. Cutting was initiated at the papillary orifice and extended cephalad along the midline of the longitudinal fold until the biliary orifice was identified. The ES was generally completed with a standard traction sphincterotome. Four patients had a sphincterotomy performed after a combined percutaneous endoscopic technique (11). Patients referred for possible SOD had sphincter of Oddi
manometry (SOM) performed in the conventional retrograde fashion using the low-compliance infusion pump system and a triple-lumen 5F catheter, as previously described (12). Each manometry catheter port was perfused at 0.25 mL/min with sterile bubble-free deionized water. The SOM was performed by the standard perfusion or the newly developed aspiration technique (13). The aspiration technique has been shown to give nearly identical manometric results to the standard technique (13) and has the advantage of reducing post-SOM pancreatic enzyme elevation and pancreatitis (14). The ductal system entered was confirmed by aspiration of clear (pancreatic) or yellow (bile) colored fluid (15). The basal sphincter pressure was defined as the baseline pressure between phasic waves (using the duodenal lumen pressure as the 0 reference point) sustained for at least 30 seconds and observed in at least two leads. Tracings were interpreted blindly by two readers, and when differences in interpretation existed, they were resolved by mutual agreement. When disagreement persisted, the reviewers’ observations were averaged. The basal sphincter pressure was considered abnormal if the mean of all observations was 2 35 mm Hg (16). The frequency, amplitude, duration, and direction of propagation of phasic waves were determined, but they were not used in assessing if the study was abnormal. Sedation for SOM was achieved with diazepam alone. All narcotics, anticholinergics, and other drugs known to affect the sphincter pressure were discontinued at least 12 hours before the SOM and avoided during it. The SOM was performed immediately before the ERCP/ES or during different days. The diagnosis of SOD was based on a clinical history of typical pancreaticobiliary pain or idiopathic pancreatitis with one or more of the following: (a) abnormal SOM results; (b) delayed contrast drainage from the common bile duct [i.e., contrast remaining in the biliary tree at 45 minutes with the patient in the supine position) and/or pancreatic duct (i.e., contrast remaining in the pancreatic duct at 9 minutes with the patient in the prone position) at ERCP; and (c) dilated pancreatic (duct diameter 2 6 mm in the head and/or 5 mm in the body) and/or common bile duct (common bile duct 2 10 mm). Patients undergoing ES were prospectively evaluated for the development of immediate complications from the procedure and observed until discharge from the hospital. Procedure-related complications were graded as mild, moderate, or severe, depending on the duration of hospitalization required to treat the complication and the need for interventional therapy (17). Post-ES pancreatitis was diagnosed when either of the pancreatic enzymes were elevated four times the upper limits of normal at 18 hours and when this elevation was associated with increased abdominal pain persisting for at least 24 hours after the procedure and requiring narcotic analgesics. Pancreatitis was graded as follows: mild, hospitalization required for outpatient admission or prolongation of planned admission to 2-3 days after the procedure; moderate, hospitalization extended 4-10 days after the procedure; and severe, hospitalization extended for > 10 days or the development of hemorrhagic pancreatitis, pseudocyst, phlegmon, or intervention required (percutaneous or surgical drainage). Bleeding was
(n oz)sl_[a~ poyq paapayed qw (ly&@m
aq01
IlaJ$0~ SEM luaped) dllexpaur paBEuEw lillnJssa33ns sEM lua!led aqd 'sglaJ3E srnoq 8p pa$3a*apQE31ug3 ‘paalq padelap E SEM aposrda alaaas Jaq$o ayJ, 'papE3 (IuaJrnD uogElnSEo3 rEIodouour PUB ‘ddelaql dexg ‘uo!$Da[u!OOO‘~JI:Iaugdauida ‘apEuodwE$ uooIIEq) lidexayl 3ldoxopua paqnbal
aSEy~~owaq
-Ed JO %v'r
Ial3E uogezgoqtua
O!yde&lo@uE
JO apostda aJaAas au0
uj paJJnxo
's)uag
ampaOoJdlsod
-EAxasuo3 paSEuEw
aJaM sggEaJwEd
ayl lal3E s&p
*sggEa~~uEd ampaDoId$sod $0~ ‘~5.0 = d)
.(sdnolEi snoaueIla+ur pue ‘Journ) ‘(s)auo~s Imp uounuo~~ pau!quxoD aql ‘SA dn0.B aOS) ~00’0 > dq ~suo~~empu~zngmds ~03examaaS,, so&/srr
(96-&l.) E'S9
(001)EZP
(98-67)E'ES (0'8) PE /98-P&)Z'S9 (Z'PZ) 09 (SG-ZZ) 8'Z9 ki'9E~ ES1 &6L-Et)L'tP (Z.6'2) 991
1WO.L
.snoauerIamg.y .JournL
(s)auWPnp~o~~o3 uowun3sLp aos
uo~~Eu1~03 67, pue &I
sluaged OM$J~$JEI 30 sapostda alaAas
‘a$ElapouI p ‘pI!tu ~1 alam aJaqL
(%p’g)sliep
.($ualajJ!p
luaJajj;rp uo
sekEp auIEs aqJu0 paurxo3lad SEM~OS
Axranbay
‘dno& aos
OZ/PZ t&/6Z oor/c9 ,PSI/ZZ
8ulpaalq ampaDoxd$sod
pue
IJJ~M s)ua;r$ed11~ *suo!snga IEmaId uourSaIqd q];[~ ampaDoJd
~Euo~~~ppEue~o3pazgE~~dsoqa~a~ ayL
z pue
@3gsgE~s 10 (%I'~)sB UaqM
JEI!UI!S~JIM palm330
SggEal3UEd
ay~u1~(sdno~8snoauE~~ax~wpue
'lourn
~0s
dq paunguo3
S~SOUiiiE!p aqJ pey
s$uagEd JO
%I'98 ‘II"JaA() '%Z'L6 u! (@I UUU s&7) pa$EAaIa SEM alnssald laJXI!qdS Ieseq aq,L '(Zgt/Lpt) paldrua$$E SEM
11 UIO~M
dIIn3ssaxns
II! sluaged JO %~'06 SBM
~~0s
‘dnoxtl (30s
II! paurro3lad aq$ %ourv
‘aUO~S~~npuour~o~pau~quro~ay~'s~dno~~(IO~a~~ u~S~~~~Ea~xxEd~~-~sod 3oh1anbaqay~~o3
~00'0 = d)
30 UogEnIEAa
x~southxp
pUE
dnokl (Ios aql II!paxnwo ‘uoy$ExIduro~ Juanbag
SggEalXIEd
JO sapos!da
~SOUr aq] SBM
S~J~JEaJXIEd
x!ayj 30 &~~pe&l
suo~~E3~~duro3 paxpur-Sz
JO addI aqL
?wlajJ!p ~wwww 10~‘cio*o < d) kh0.d SB +I0
= d)suog
0) JSEJJUOLI U! s)uagEd
‘(SaJEp!pUED IE@JnS
103 holo~a~xr!qds
uoyyasu! lua)s ~03 arnj3nls dreg!q qg~
ay$ JO O/o~'g~‘JaAaMOH *sjuagedjO oho'p U! %~.I.m330 *p a1qEL uo palElnqe$ ale &aAas aql pue
p~epUE~SUE~~(%z'g)uo!~le~~Id~o3EU~~InSa~O~dIay~I alouIouse~duIo~ola~~u!~ds~n3a~d'(600'0 -EXpUy Jay10 IIT? “03 %L’IT
aos
JO %r'rz u! paqnbal SEM $1 '(E a[qxL) SwatlEd JO %f,'sr uy paurxo3xad SCM huolola~wgds JnDaxd IOU
‘so.0 < d) (%s*II)
uo JO (%z’o~)sz ~0s
sdep w=yj!p
ac[l aaoJaq dIa~Erpauuu! paurxo3
uayM
huanbaq
sy$EaJxIEd
Iward
sggearxred
3p101q3 ‘(p)auro;rpudsdwns
‘(~)ddE~aq$
'(PC) suoseal snoaue~~aw~ur (p) pue
!(~)loU~n] hIlndure
pue ‘(Et)Eu.ro~dti-[10 euroup
(3)
ggt)aos
(e)(r aIqEJ,)a.IaMhuo~o~a~wryds
ay] ~03 suogexpu!
aqL
*slua;r)ed czf, uo paurlojlad
pue 6862
SEM SB ‘0665 wnsnv
q~wv
uaawaa
Ienba ~J!M palm3
‘dno&i (Ios ayl XII s$uagEd
pa~~nba~uogex~dwo~ $0~ ‘Z~~YJ
~03 0Qj.z HUE
ayl
ay$ JO auoN = d)
SUO~JEX~U! ~IIE~!J~!JE~S
‘(~ua~agp
%&‘E
‘(gp()‘O
= d)
SaUOJS
JO3 %6-v 01 $SEQUO3 KIT%8'()t uy paJm330
10~ ~3 lage suogE3gduro3
a[os
*(fj()o*o = d tsdnoh'snoau
‘Jotunl ‘(s)auolsIx-tp uoun1103 pautq 103 pawo3lad
JaqJO dUE
SEM ~3 uayM
suoyExlduro3
dpuanba;r3
‘JaAaMOH
Sz laJJE a)EJ uogexIduro3
~~~dno&iaos
paUIaap IOU sluayed) (P)
~~uo~y~‘(~)a~a~o~~opa~o~~‘(~)uo~~.~asu~s~saq~so~dop -Ua “03 yEaI aI!q ‘(01) IEAOUIaJ aUO$S 3gEa;rXIEd JO3 luajs ~o/pue uoge~tp ~03 s@!h1~~oq3 Su!soraps hxu -!Jd kpnpuy
-183 3!~E~SE~a~‘(OZ)EuIOU!~~~O~~UEIOq~‘(pZ)E~OU!~ -;re~~!~ea~3uEdSu!pnI~u;r ‘(09)(uo~~~asu~s~saq~so~dop !(cgT)(s)auols 13np uounuo3 (q) -ua ~03) JownJ
t(s$uayEd QIEWYEP -lad SEM
-30 suogEx~duro3 iikrou_rv *ha%ns
*(luaJajjrp QE~+~)EJS snoaue~~ax~w
IOU '6L()'O= d) SJOUInJ “03
hy~q
-ElIax;rurpue
-III03ayi “03 %E’~T 'SA %Q'O~) UOnEXpU! ~03 ueq~ aos
a.IOUIpa.Im~~O Oj6.9 SEM
'(7,aIqEL)
IIElaAO aqJ,
‘(LOO*0 > d !uauroM yo8*zfj)axreururopald aIEura3 paynxu pue
(roo*f~ > d:s~Iead~.p~‘uEaur)a8e~a~unoddq(~a~qE~) pazpalxsey3
ys~guo3u1~(~~u.1u.10~>
ayl ‘sdno&llaq$o aylo)
o~B~u1u1~~7
s~~a~nssa~d
ayl ‘s)uagEd 30 Oh8.P IEUO!$~~E UE
la$Bu!qds IWEq
.palmbar hra%ns JO q30ys xldas ‘alaaas pm !uoguaMalu! snoaue~nmad JO ~!do~sopua~odderarIlIel!dsoq30 sleepE< Bu!~nba~ssauIl! ggdas JO aplqa3 ‘almapour !ampaDold aql lal3e sleep E 5 103 3$3& < amlmadmal ‘p~pu se papeti SBM uog3a3uI *pa+nbaJ (p=?@ms JO snoauelnxad) uo!lualwalu! JO sleep 1~17 103 dde -lay4 lexpam ‘alahas pm :shp 01-t ~03 @cnpatu paleaq uogeJo3Jad algap due ‘awapotu !shp E r ~03 uogms pm spy3 dq paleaq 1seguo3 ~0 pyj 30 yz?aI @!!Is haa Quo JO alq!ssod‘P1!":SMOIIOJse papeti Sl?Muoyxo3.Iad jpz$hs JO 3yde~Bo@he) pal;mbal uo;rwahIalur JO sIIa3 pal paqxd 30 n 5 z 30 uo!sn3sueq ‘aJaaas pm !pagnbar hafhns JO Uo~~uaAJa~u~ 3ydeJSo@Ie ou ql;[~‘SIla:, poolq pal paqxd 30 n fF JO uo!sn3sueJJ ‘aJeJapouI :paJFnbaJ uo!sysue~ 0U qJ!M ‘8 E> 30 IaAaI u!qo@omaq UT aseamap e ~J!M uo!~e!~osseu;r s!sauralewaq ~0 malam se qms %npaaIq
30 amapyaa
U~!i!!HUIUI(-J~~ amssa~d~a~~u~yds~EseqEpEq~o~~~~)
(sydomopua
IOU ‘.a.!) pxy.up
t ‘ON ‘LOL ‘IOA 63070EI.L.N30XLSV3
‘p~itu se papeti
‘TV 13 mms
OLOI
October
COMPLICATIONS
1991
Table 2. Endoscopic
Sphincterotomy-Znduced
Complications Complications
Indication
No. of ES
SOD Common duct stone(s) Tumor Miscellaneous
166 163 60 34
Total
423
and 48-Hour
OF ENDOSCOPIC
SPHINCTEROTOMY
and 30-Day Mortality
Surgery
48-h mortality
39day
mortality
(%I”
(%Jb
(%)I
10.8d 4.9 3.3 2.9
0 0 0 0
0 0 0 0
0.0, 0.6 6.7 5.9
6.9
0
0
1.7
(%I
“Refers to the percentage of patients requiring surgery for the complication. bRefers to the percentage of patients dying within 48 hours of the ES. “Seven patients were lost to follow-up and were excluded from tabulation, Six of the seven deaths were seemingly sphincterotomy, whereas one was possibly related. dP = 0.009,SOD group vs. the combined common duct stone, tumor, and miscellaneous groups. "P = 0.03,SOD group vs. combined common duct stone, tumor, and miscellaneous groups.
and treatment with a Bicap hemostatic probe (CirconACMI, Stamford, CT) on two occasions (the first session with Bicap controlled the bleeding for only 2 days). Angiography failed to identify a bleeding source. Four episodes of infection (0.9% of patients) occurred. Two patients experienced cholangitis 8 and 25 days after ES and stent insertion for pancreatic carcinoma. Both patients were managed successfully with antibiotics and stent change. One SOD patient had transient low-grade fever and abnormal hepatic chemistries after a precut ES. This patient improved on IV antibiotics and was discharged from the hospital 4 days later. An 86-year-old nursing home patient died of probable pancreatic sepsis 14 days after sphincterotomy and 7 days after extraction of pancreatic duct stones. She had been clinically stable for 13 days after the ES. Autopsy was denied. Although the exact source for her terminal febrile illness was not clinically apparent, we are tallying this as an ESrelated mortality. Perforation occurred in two patients (0.5%). One patient was managed with nasogastric suction, nasobiliary tube drainage, and gut rest, whereas the other was successfully treated by nasogastric suction and
Table 3. Early Complications
1071
of Precut Sphincterotomy Complications
Precut ES
Precut
No precut
Indication
No. of ES
(%I0
(%I
(%I
SOD All othefl
166 257
21.1 11.7'
8.6 3.3d
11.5 4.4'
Total
423
15.4
6.2
7.0
NOTE. Precut sphincterotomy was performed over a previously placed pancreatic stent in 63% of precut cases. “Refers to the percentage of patients undergoing sphincterotomy who required the precut technique. ‘Combines the common duct stone, tumor, and miscellaneous groups. "P = 0.009;dP > 0.05;"P = 0.012.
unrelated
to the
gut rest alone. No episodes of basket impaction or acute cholecystitis occurred in this series. Complications after ES for all indications occurred in 11 .O% of patients with a normal-diameter common bile duct (< 10 mm) in contrast to only 2.8% of patients with a dilated biliary tree (P < 0.001) (Table 5). Similarly, more SOD patients had a complication when ES was performed in the presence of a normaldiameter common bile duct (15.1% vs. 3.3%; P = 0.020). The relationship of the common bile duct diameter and the distribution of complications after ES for all patients (combined group) and for SOD patients is shown in Figures 1 and 2, respectively. For the combined group, complications occurred in 19.6% of patients with a small-diameter common bile duct (I 5 mm) in contrast to 8.2% with an intermediate size (6-< 10 mm) (P = 0.023) and 2.8% with a dilated common bile duct (210 mm) (P < 0.001 vs. small diameter and P = 0.02 vs. intermediate size). Among the SOD patients, 37.5% with a small-diameter common bile duct had a complication in contrast to 8.5% with an intermediate size (P < 0.001) and 3.3% with a dilated common bile duct (P < 0.001 vs. smalldiameter common bile duct; P > 0.05 for intermediate size vs. dilated common bile duct). Thirty-day follow-up was obtained in 98.3% of patients (Table 2). The overall mortality rate during this period was 1.7%. The mortality rate for the SOD group was 0% (P = 0.03 for the SOD group vs. the combined common duct stone(s), tumor, and miscellaneous groups) in contrast to 0.6% (P > 0.05, not statistically different), 6.7% (P < O.OOl), and 5.9% (P = 0.002) for the common duct stone, tumor, and miscellaneous groups, respectively. Six of the seven deaths were seemingly unrelated to the ES. One patient died of arrhythmia 11 days after ES and 7 days after surgical resection of a large choledochocele. Four patients with a malignancy died 12, 14, 21, and 29 days after the ES and stent placement of metastatic cancer, urosepsis, aspiration pneumonia, and metastatic cancer, respectively. One patient died of myocar-
GASTROENTEROLOGY
1072 SHFdWiANETAL.
Vol. 101, No. 4
Table 4. Type and Grade of Complications After Endoscopic Sphincterotomy Pancreatitis
Bleed Indication
Mld
Mod
Sev
SOD Common duct stone(s) Tumor” Miscellaneous”
1 0 0 0
2 1 0 0
0 2 0 0
Total
1
3
2
NOTE. See Materials “See text for specific
and Methods indications.
for definition
Sev
Mld
Mod
Sev
Mld
Mod
Sev
7 4 0 0
4 0 0 0
2 0 0 0
0 0 0 0
1 0 2 0
0 0 0 1
0 0 0 0
1 1 0 0
0 0 0 0
11
4
2
0
3
1
0
2
0
of grade of complication
(Mld, mild; Mod, moderate;
Endoscopic sphincterotomy may be an effective therapy for patients with SOD (l-4). Despite its potential use for this disease, ES has been reported to be associated with a higher complication rate when performed for SOD than when performed for choledocholithiasis (Table 6). Among the 51 patients with SOD treated by Thatcher et al. (Z), 8 complications (15.7% of patients) occurred in 7 patients (4 patients with perforations, 2 with hemorrhage, and 2 with pancreatitis). Complications were more likely to occur in the presence of a nondilated common bile duct. Surgery was required in 3 patients with perforations, but no deaths occurred. In the series reported by Neoptolemos et al. (3), 10 complications in 8 patients occurred among the 32 attempted sphincterotomies. These complications were bleeding in 3, acute pancreatitis in 4, perforation in 2, and acute cholangitis in 1 patient. Three patients required surgery, but no deaths occurred. Similarly, Leese et al. (5) reported that complications were more common when ES was performed for papillary stenosis (6 of 37 patients, 16.2%) than for common bile duct calculi (33 of 319, 10.3%). In contrast, Roberts-Thomson and Toouli (4) reported only 2 (4%) complications (bleeding and acute pancreatitis) in 50 patients treated for SOD.
Complications: Relation to Bile Duct Diameter Frequency
of CBD
Our entire
series
Sev, severe).
Geenen et al. also found the complication rate for patients with SOD (5 of 80 patients, 6.3%) to be similar to that of patients with common bile duct stones (18). In Gregg’s (19) series of 120 patients with SOD, 13 cases of pancreatitis (2 hemorrhagic), 2 episodes of bleeding, and 1 perforation occurred after ES for SOD. It has been postulated that the hypertrophic or fibrotic sphincter in patients with SOD increases the risk of pancreatitis, because extended periods of high intensity current may be required to sever the sphincter. Duodenal perforation has been reported to occur more frequently in SOD patients (19,20). Among the four perforations reported by Thatcher et al., three had a nondilated common bile duct (2). The increased risk of perforation may be a reflection of the relatively small duct size lowering the safety margin in performing a complete ES (9). Classen found a mortality of 2.2% among 813 patients undergoing ES for suspected SOD in contrast to 1% mortality among 7585 patients undergoing ES for stone disease (8). Most studies reporting the complication rate of ES for SOD are limited by their retrospective nature and use heterogenous populations and a small number of patients. Our study was undertaken to prospectively evaluate the ES-induced complication rate in a large group of patients with manometrically documented SOD in the majority (86.1% had abnormal basal sphincter pressures) and to compare it with the complication rate for other indications. The complication rate after ES was significantly higher for patients
Discussion
Diameter
Perforation
Mod
Mld
dial infarction 5 days after successful removal of common duct stones. As noted, one patient died of probable pancreatic sepsis possibly related to the sphincterotomy performed 14 days before.
Table 5. Endoscopic Sphincterotomy-Induced
Infection
Biliary
of complications
stones
SOD
Tumor
Miscellaneous
lOmm
23/210 (11.0%)” 6/213 (2.8%)b
5/48 (10.4%) 3/115 (2.6%)d
16/106 (15.1%)” 2/60 (3.3%)’
2/38 (5.3%)g o/22 (O.O%)h
O/18 (0.0%)’
Total
29/423
8/163 (4.9%)
18/166
2/60 (3.3%)
l/34 12.9%)
(6.9%)
NOTE. Numbers represent the frequency of complications defined as dilated. ’ VS. bP < 0.001; ’ VS. dP = 0.035; ’vs. rP = 0.02; gvs. * and
for the given
range
’vs. ‘not statistically
(10.8%)
of common different,
bile duct
P > 0.05.
(CBD) diameter;
l/16(6.3%)'
CBD 2 10 mm was
October 1991
COMPLICATIONS
100
0
86
UI E 0
All
patients
? ?Complicstlons
90
I
I
i
70 60 1
% P
50
P
40 1
30
-I
20 10 0
1
<5
6-7
0-9
Common
10-11
Bile Duct Diameter
12-13
>
14
(mm)
Figure 1. Frequency of complications after ES for all indications: relation to common bile duct diameter [a, small (15 mm); b, intermediate (6- < 10 mm); c, dilated (2 10 mm)]. 0, All patients; ?,? complications. Percentages represent the frequency of complications for the given common bile duct diameter. a vs. b, P = 0.023;a vs.c, P < 0.001;b vs. c, P = 0.020.
with SOD than for other indications (10.8% vs. 4.3% for all other indications combined). In the SOD group, pancreatitis occurred in 7.8%, hemorrhage in 1.8%, perforation in 0.6% and infection in 0.6% of patients. In contrast, in the combined common duct stone, tumor, and miscellaneous group, pancreatitis occurred in 1 .S%, bleeding in 1.2%, perforation in 0.4%, and infection in 1.2% of patients. Our impression is that the common bile duct is more difficult to deeply and selectively cannulate in SOD patients. This may result in repeated cannulation attempts and pancreatic duct injections, which have been shown to result in a higher frequency of postprocedure pancreatitis (6,21,22). Papillary edema and sphincter of Oddi spasm resulting from mechani48
40
P
24
P 16
25
6-7 Common
8-9
10-11
Bite Duct Diameter
12-13
2
14
(mm)
Figure 2. Frequency of complications after ES for SOD: relation to common bile duct diameter [a,small (I 5 mm); b, intermediate (6- < 10 mm): c, dilated (2 10 mm)]. Cl, All patients; ? ? complications. Percentages represent the frequency of complications for the given common bile duct diameter. a vs. b, P < 0.001; a vs. c, P < 0.001;b vs. c, not statistically different (P < 0.05).
OF ENDOSCOPIC SPHINCTEROTOMY
1073
cal trauma may restrict the flow of pancreatic juice and result in pancreatic edema and pancreatitis (23). In patients with common duct stones, cannulation is often easier and pancreatography frequently not required, thus limiting mechanical trauma to the papilla. The significantly higher need for the precut technique in patients with SOD in this series supports the notion that selective cannulation of the bile duct is often difficult in these patients. The incidence of pancreatitis after SOM may be as high as 20% (14,24) and, theoretically, might contribute to the risk of this complication if performed immediately before the ES. However, when SOM was performed before ES on the same day or on different days (or not at all), the incidence of all complications and of pancreatitis was no different. However, our routine use of the SOM aspirating technique almost certainly decreased the frequency of SOM-associated pancreatitis (14). The overall complication rate in the present series was 6.9%. No patient required surgical intervention, and no immediate mortality occurred. These rates for very serious complications are somewhat lower than those reported in the literature (22,25-27). The current series seemed to have fewer serious complications compared with what was informally recalled from our initial lo-year experience in which a few cases of post-ES phlegmon, abscess, pseudocyst, or bleeding resulted in death or laparotomy. In part, our more recent complication rates may be related to greater operator experience, more attention to cannulation/ES details, and our younger (mean age, 55.3 years), relatively more fit population. A variety of precautionary measures are used in an effort to avoid complications. We strive to always assure drainage of an obstructed duct, because the risk for cholangitis is substantially higher if the duct is not decompressed (28). In the present series, nine patients had biliary stents placed because of failure to clear the common duct of stones post-ES (stones were either large or proximal to a narrowing). In five patients (8.3%) with a malignancy, drainage took place percutaneously after unsuccessful placement of stents endoscopically. These interventions almost certainly reduced the incidence of postprocedure cholangitis. Moreover, using alternative methods to fragment common duct stones (e.g., laser and mechanical lithotripsy) and placing long-term biliary stents when the common duct is not cleared of stones makes large sphincterotomies (with the associated increased risk of complications) often unnecessary. Prophylactic antibiotic therapy is continued after the procedure if an obstructed pancreatic or biliary tree is found at ERCP (29). We attempt to reduce the risk of hemorrhage by correcting a coagulopathy, discontinuing (ideally) aspirin and nonsteroidal antiinflammatory use 7-10 days before the ES, limiting the length of the
1074
SHERMAN ET AL.
GASTROENTEROLOGY Vol. 101. No. 4
Table 6. Reported Complication Rates Associated With Endoscopic
Sphincterotomy
for Sphincter of Oddi Dysfunction
No. of complications Study Thatcher et al. (2) Roberts-Thomson et al. (4) Neoptolemos et al. (3) Leese et al. (5) Geenen et al. (18)
n 51 50 32 37
Gwg(l9)
80 120
Total
370
Total
Bleed
8" 2 10b 6
2 1 3 _c
5 16
2
Pant
Perf
Chol
No. of surgeries
No. of deaths
2 1 4
4 0 2
0 0 1 -
3 0 3 -
0 0 0 -
13
0
-
1
0 -
-
47(12.70/b)
Pant, pancreatitis; Perf, perforation; Chol, cholangitis. “Eight complications occurred in seven patients. *Ten complications occurred in eight patients. “Data unavailable.
cut to the amount required, and using coagulation current (alternating with blended current) for the cephalad half of the incision or any recut endoscopic sphincterotomy. If bleeding is significant, we attempt to control it endoscopically. If endoscopic therapy fails to control the hemorrhage, angiographic embolization is quickly attempted (30). We attempt to reduce the risk of pancreatitis by limiting cannulation time; avoiding the number of injections, acinarization, and coagulation current near the pancreatic duct orifice; meticulously disinfecting endoscopes; using low osmolarity agents in patients with a prior episode of pancreatitis (or prior episode of ERCP-associated pancreatitis); premeditating with oral steroids and antihistamines in patients with previous contrast media reactions; and using the aspirating SOM catheter when SOM of the pancreatic duct sphincter is performed (22). When the endoscopist is having difficulty achieving deep biliary cannulation, a choice must be made whether to continue cannulation attempts, terminate the procedure, perform precut sphincterotomy, use percutaneous assistance, or consider surgery. Multiple factors must be considered in this choice. We tend to use precut techniques fairly aggressively. Precut ES was used in 15.4% of patients in this series. This technique has been reported to be potentially hazardous, particularly when performed by inexperienced endoscopists (3 1). However, our complication rate for precut ES was not statistically different from standard ES and is in agreement with the complication rate reported by Dowsett et al. (10) for this technique. Actually, a trend toward greater safety was observed from precutting. This may, in part, be related to the pancreatic stent placed before precutting or to the fact that additional prolonged sphincter probing was often avoided. Endoscopic sphincterotomy-induced complications occurred in 11.0% of patients with a normaldiameter common bile duct and was significantly
reduced in the presence of a dilated common bile duct. These results are in support of those of Mustard et al. (2O), Thatcher et al. (2) and Gregg (19). Patients with small-diameter (I 5 mm) common bile ducts had significantly more complications than patients with intermediate-size (6 to < 10 mm) or dilated (2 10 mm) common bile ducts. Moreover, patients with intermediate-size ducts had more adverse events after ES than patients with a dilated common bile duct. Although this separation of normal-sized common bile ducts into small and intermediate is somewhat arbitrary, it serves to emphasize the significant rate of complications for SOD patients with a small-diameter duct (37.5%) and the stepwise reduction in adverse events for those with an intermediate size and a dilated common bile duct, respectively. Thirty-day mortality was 1.7%. However, the procedure-related mortality was believed to be 0.2%. Again, these procedure-induced fatalities are lower than many reports in the literature and may be related to our younger, lower-risk population and attention to details of technique. Our data again support the impression that ES for SOD is more hazardous than for other indications. Because of the relatively high incidence of major complications, some authorities recommend a trial of medical therapy before attempts at endoscopic treatment (19). ES for SOD has been reported to result in pain relief in 50%-90% of patients (l-4). Because surgical sphincteroplasty with septoplasty severs the pancreatic and biliary sphincters, would surgical therapy for these patients be preferable? A prospective randomized study addressing the therapeutic benefits and complication rates of endoscopic vs. surgical therapy is in progress at our institution. This study will help detect the frequency of restenosis of the sphincterotomy site, a factor not addressed in this manuscript. In summary, ES for SOD is associated with an increased risk of complications and more frequently
October
1991
requires the precut technique. Compared with standard ES, precut ES (usually with stenting) was not associated with an increased frequency of complications. The risks of complications after ES appears to be less for those with an intermediate or dilated common bile duct. More importantly, the risk of a complication is considerable for a small-diameter common bile duct, particularly when ES is performed for SOD.
References 1. Geenen
JE, Hogan WJ, Dodds WJ, Toouli J, Venu RP. The efficacy of endoscopic sphincterotomy after cholecystectomy in patients with sphincter of Oddi dysfunction. N Engl J Med 1989;320:82-87. 2. Thatcher BS, Sivak MV, Tedesco FJ, Vennes JA, Hutton SW, Achkar EA. Endoscopic sphincterotomy for suspected dysfunction of the sphincter of Oddi. Gastrointest Endosc 1987;33:9195. 3. Neoptolemos JP, Bailey IS, &r-Locke DL. Sphincter of Oddi dysfunction: results of treatment by endoscopic sphincterotomy. Br J Surg 1988;75:454459. IC, Toouli J. Is endoscopic sphincterotomy 4. Roberts-Thomson for disabling biliary-type pain after cholecystectomy effective? Gastrointest Endosc 1985;31:370-373. 5. Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their management following endoscopic sphincterotomy: results in 394 consecutive patients from a single centre. Br J Surg 1985;72:215-219. 6. Podolsky I, Haber GB, Kortan P, Gray R. Risk factors for pancreatitis following ERCP: a prospective study (abstr). Am J Gastroenterol 1987;82:97A. 7. Siegel JH. Endoscopic papillotomy in the treatment of biliary tract disease: 258 procedures and results. Dig Dis Sci 1981;26: 1057-1064. 8. Classen M. Endoscopic papillotomy-new indications, shortand long-term results. Clin Gastroenterol 1986;15:457-469. 9. Krims PE, Cotton PE. Papillotomy and functional disorders of the sphincter of Oddi. Endoscopy 1988;20:203-206. 10. Dowsett JF, Polydorou AA, Vaira D, D’Anna LM, Ashraf M, Croker J, Salmon PR, Russell RLG, Hatfield ARW. Needle knife papillotomy: how safe and how effective? Gut 1990;31:905908. 11. Dowsett JF, Vaira D, Hatfield ARW, Cairns SR, Polydorou A, Frost R, Croker J, Cotton PB, Russell RLG, Mason RR. Endoscopic biliary therapy using the combined percutaneous and endoscopic techniques. Gastroenterology 1989;96:1180-1186. 12. Geenen JE, Hogan WJ, Dodds WJ, Stewart ET, Arndorfer RC. Intraluminal pressure recording from the human sphincter of Oddi. Gastroenterology 1980;78:317-324. 13. Sherman S, Troiano FP, Hawes RH, Lehman GA. Does continuous aspiration from an end and side port in a sphincter of Oddi manometry catheter alter recorded pressures? Gastrointest Endosc 1990;36:500-503. 14. Sherman S, Troiano FP, Hawes RH, Lehman GA. Sphincter of Oddi manometry: decreased risk of clinical pancreatitis with use of a modified aspirating catheter. Gastrointest Endosc 1990;36:462-466. 15. Carr-Locke DL, Gregg JA. Endoscopic manometry of pancreatic
COMPLICATIONS
OF ENDOSCOPIC SPHINCTEROTOMY
1075
and biliary sphincter zones in man: basal results in healthy volunteers. Dig Dis Sci 1981;26:7-15. 16. Guelrud M, Mendoza S, Rossiter G, Villegas MI. Sphincter of Oddi manometry in healthy volunteers. Dig Dis Sci 1990:35:3840. comparisons and confusion: com17. Cotton PB. Complications, mentary. In: Cotton PB, Tygat GNJ, Williams CB, eds. Annual of gastrointestinal endoscopy, 1990. London: Current Science Ltd., 1990:7-g. 18. Geenen JE, Hogan WJ, Dodds WJ. Sphincter of Oddi. In: Sivak MV, ed. Gastroenterologic endoscopy. Philadelphia: Saunders, 1987:735-751. 19. Gregg JA. Function and dysfunction of the sphincter of Oddi. In: Jacobson IM, ed. ERCP: diagnostic and therapeutic applications. New York: Elsevier Science, 1989:139-170. 20. Mustard R, Mackenzie R, Jamieson C, Haber GB. Surgical complications of endoscopic sphincterotomy. Can J Surg 1984; 27:215-217. 21. Hamilton I, Lintott DJ, Rothwell J, Axon ATR. Acute pancreatitis following endoscopic retrograde cholangiopancreatography. CIin Radio1 1983;34:543-546. 22. Sherman S, Lehman GA, ERCP- and endoscopic sphincterotomy-induced pancreatitis. Pancreas 1991;6:350-367. Nordenstam CG, Brackett 23. Saari A, Kivisaari L, Standertskjold, K, Schroder T. Experimental pancreatography: a comparison of three contrast media. Stand J Gastroenterol 1988;23:53-58. 24. Rolny P, Anderberg B, Ihse I, Lindstrom E, Olaison G, Arvill A. Pancreatitis after sphincter of Oddi manometry. Gut 1990;31: 821-824. 25. Tedesco FJ, Vennes JA, Dreyer M. Endoscopic sphincterotomy: the USA experience. In: Okabe H, Honda T, Oshiba F, eds. Endoscopic surgery. New York: Elsevier Science, 1984:41-46. 26. Vaira D, D’Anna L, Ainley C, Dowsett J, Williams S, Baillie J, Cairns S, Croker J, Salmon P, Cotton P. Russell C, Hatfield A. Endoscopic sphincterotomy in 1000 consecutive patients. Lancet 1989;2:431433. 27. Geenen JE, Vennes JA, Silvis SE. Resume of a seminar on endoscopic sphincterotomy (ERS) Gastrointest Endosc 1981;27: 31-38. 28. Ostroff JW, Shapiro HA. Complications of endoscopic sphincterotomy. In: Jacobson IM, ed. ERCP: diagnostic and therapeutic applications. New York: Elsevier Science, 1989:61-73. 29. Hershey SD, Sugawa C, Cushing R, Ledgerwood AM, Lucas CE. The value of prophylactic antibiotic therapy during endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1982;155:801-803. 30. Saeed M, Kadir S, Kaufman SL, Murray RR, Milligan F, Cotton PB. Bleeding following endoscopic sphincterotomy: angiographic management by transcatheter embolization. Gastrointest Endosc 1989;35:300-303. risky technique for experts 31. Cotton PB. Precut papillotomy-a only. Gastrointest Endosc 1989;35:578-579.
Received October 24,199O. Accepted February 4,199l. Address requests for reprints to: Glen A. Lehman, M.D., Indiana University Hospital, 926 West Michigan Street, Room N541, Indianapolis, Indiana 46202. Dr. Stuart Sherman was the recipient of an American Society for Gastrointestinal Endoscopy/Olympus Corporation of America Tier III scholarship.
Complications of Endoscopic Sphincterotomy A Prospective Series With Emphasis on the Increased Risk Associated With Sphincter of Oddi Dysfunction and Nondilated Bile Ducts STUART
SHERMAN,
THOMAS
A. RUFFOLO,
ROBERT
H. HAWES,
and GLEN A. LEHMAN Division of Gastroenterology/Hepatology, Indiana University Hospital, Indianapolis, Indiana
Mostly retrospective series with limited use of sphincter of Oddi manometry have indicated that early complications are more common when endoscopic sphincterotomy is performed for sphincter of Oddi dysfunction than for common duct stones. The current study was undertaken to prospectively evaluate the frequency and type of complications of endoscopic sphincterotomy performed for sphincter of Oddi dysfunction compared with endoscopic sphincterotomy performed for other conditions. Four hundred twenty-three patients underwent sphincterotomy for sphincter of Oddi dysfunction (166), common duct stone(s) (163),tumor (60),and miscellaneous reasons (34). Patients were observed in the hospital for at least 24 hours after the procedure, and 36-day follow-up data were obtained. The overall complication rate was 6.9%, but complications were more frequent when sphincterotomy was performed for sphincter of Oddi dysfunction than for all other indications (10.6% vs. 4.3%; P = 0.009). Precut sphincterotomy was more frequently required in the sphincter of Oddi dysfunction group (21.1% vs. 11.7%, P = 0.009)but was no more likely to result in a complication (6.2%)than standard sphincterotomy. The risk of a complication was considerable for a small-diameter common bile duct (15 mm), particularly when sphincterotomy was performed for sphincter of Oddi dysfunction (37.5%). The overall 30-day mortality rate was l.7%, but the procedure-related mortality rate was believed to be 0.2%. Itis concluded that endoscopic sphincterotomy for sphincter of Oddi dysfunction is more hazardous than for other conditions, particularly when a small common bile duct is present.
ndoscopic sphincterotomy (ES) is gaining acceptance as an effective therapy for patients with sphincter of Oddi dysfunction (SOD) (l-4). However, reports have indicated that early complications are more common with ES performed for SOD than for Additionally, the morcommon duct stones (2,3,5-g). tality rate (two times higher for SOD) and late stenosis rate after ES (approximately four times higher for SOD) are substantially higher in patients with SOD (8,9). Most of these data come from retrospective series involving small numbers of patients with limited use of manometry to document SOD. The current study was undertaken to prospectively evaluate the frequency and type of complications of ES performed for SOD compared with ES performed for other indications.
E
Materials
and Methods
Between March 1989 and August 1990, endoscopic retrograde cholangiopancreatography (ERCPs) was performed 1204 times at Indiana University Hospital. The subjects of the study were the 423 (35.1%) patients who also underwent endoscopic biliary sphincterotomy. Four hundred ten (97%) were hospitalized and observed for complications for at least 24 hours after the procedure. Seven patients were transferred back to their referring hospital after ES, whereas the remaining 6 were carefully observed
Abbreviations used in thispaper: ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy; SOD, sphincter of Oddi dysfunction; SOM, sphincter of Oddi manometry. o 1991 by the American Gastroenterological Association 0016-5085/91/$3.00
October 1991
COMPLICATIONS
OF ENDOSCOPIC SPHINCTEROTOMY
1069
for at least 6 hours in an outpatient setting and contacted by telephone follow-up. Thirty-day follow-up data were obtained by chart review or telephone interview. Patients who underwent pancreatic sphincterotomy of the major or minor papilla are not included in this series. Patients underwent preendoscopic assessment including hepatic and pancreatic chemistries, complete blood count, and coagulation studies. The prothrombin time was within 3 seconds of the control value or corrected to that time with vitamin K and/or fresh frozen plasma before the procedure in all patients. Tests for bleeding times were not performed routinely; however, they were performed and the bleeding time was required to be normal if the patient had taken platelet-affecting drugs. A platelet count of >5O,OOO/~L was required before ES. Patients were asked to discontinue use of salicylates and other platelet-affecting drugs 7 days before [in nonemergent cases] the ES and for 14 days after the ES. All patients received prophylactic antibiotics (cefoperazone was generally used unless the patient was allergic to cephalosporins) before the procedure (and for at least 24 hours after the procedure, if an obstructed biliary tree was found at ERCP). Patients were administered pharyngeal local anesthesia and sedated with IV diazepam and Demerol (Meperidine; Wyeth Laboratories, Philadelphia, PA) as needed. Duodenal relaxation was obtained with glucagon. Generally, 30% meglumine diatrizoate was used for cholangiography, whereas a 60% concentration was used for pancreatography. Full pancreatography was performed in SOD patients, but, often, in biliary stone patients, the pancreatic duct was only (intentionally) partially filled. Endoscopic sphincterotomy was performed using Olympus duodenoscopes (Olympus Corporation of America, Lake Success, NY) and Erlangen pull-type sphincterotomes. Sphincterotomes with variable nose and cutting wire lengths, as well as those used with and without guidewires, were used interchangeably and tallied together. A controlled cut was made in stepwise fashion using short pulses of blended current for the lower half of the incision and alternating blended and pure coagulation current for the cephalad half. The length of the incision was somewhat tailored to the indication for the sphincterotomy and the size of the common duct stone(s) if present. The cut was extended through 75% of the longitudinal fold in all patients and through 85%-90% of the fold in most patients. The entire longitudinal fold was cut if this fold was short ( < 10 mm) or if biliary stones > 10 mm were present. Sphincterotomies were performed by three experienced endoscopists (S.S., R.H., and G.L.). When selective cannulation of the common bile duct with the standard sphincterotome was not achieved, a precut sphincterotomy was performed using a needle knife (lo), usually (in 63% of patients) over a previously placed 5-7F pancreatic stent. Cutting was initiated at the papillary orifice and extended cephalad along the midline of the longitudinal fold until the biliary orifice was identified. The ES was generally completed with a standard traction sphincterotome. Four patients had a sphincterotomy performed after a combined percutaneous endoscopic technique (11). Patients referred for possible SOD had sphincter of Oddi
manometry (SOM) performed in the conventional retrograde fashion using the low-compliance infusion pump system and a triple-lumen 5F catheter, as previously described (12). Each manometry catheter port was perfused at 0.25 mL/min with sterile bubble-free deionized water. The SOM was performed by the standard perfusion or the newly developed aspiration technique (13). The aspiration technique has been shown to give nearly identical manometric results to the standard technique (13) and has the advantage of reducing post-SOM pancreatic enzyme elevation and pancreatitis (14). The ductal system entered was confirmed by aspiration of clear (pancreatic) or yellow (bile) colored fluid (15). The basal sphincter pressure was defined as the baseline pressure between phasic waves (using the duodenal lumen pressure as the 0 reference point) sustained for at least 30 seconds and observed in at least two leads. Tracings were interpreted blindly by two readers, and when differences in interpretation existed, they were resolved by mutual agreement. When disagreement persisted, the reviewers’ observations were averaged. The basal sphincter pressure was considered abnormal if the mean of all observations was 2 35 mm Hg (16). The frequency, amplitude, duration, and direction of propagation of phasic waves were determined, but they were not used in assessing if the study was abnormal. Sedation for SOM was achieved with diazepam alone. All narcotics, anticholinergics, and other drugs known to affect the sphincter pressure were discontinued at least 12 hours before the SOM and avoided during it. The SOM was performed immediately before the ERCP/ES or during different days. The diagnosis of SOD was based on a clinical history of typical pancreaticobiliary pain or idiopathic pancreatitis with one or more of the following: (a) abnormal SOM results; (b) delayed contrast drainage from the common bile duct [i.e., contrast remaining in the biliary tree at 45 minutes with the patient in the supine position) and/or pancreatic duct (i.e., contrast remaining in the pancreatic duct at 9 minutes with the patient in the prone position) at ERCP; and (c) dilated pancreatic (duct diameter 2 6 mm in the head and/or 5 mm in the body) and/or common bile duct (common bile duct 2 10 mm). Patients undergoing ES were prospectively evaluated for the development of immediate complications from the procedure and observed until discharge from the hospital. Procedure-related complications were graded as mild, moderate, or severe, depending on the duration of hospitalization required to treat the complication and the need for interventional therapy (17). Post-ES pancreatitis was diagnosed when either of the pancreatic enzymes were elevated four times the upper limits of normal at 18 hours and when this elevation was associated with increased abdominal pain persisting for at least 24 hours after the procedure and requiring narcotic analgesics. Pancreatitis was graded as follows: mild, hospitalization required for outpatient admission or prolongation of planned admission to 2-3 days after the procedure; moderate, hospitalization extended 4-10 days after the procedure; and severe, hospitalization extended for > 10 days or the development of hemorrhagic pancreatitis, pseudocyst, phlegmon, or intervention required (percutaneous or surgical drainage). Bleeding was
(n oz)sl_[a~ poyq paapayed qw (ly&@m
aq01
IlaJ$0~ SEM luaped) dllexpaur paBEuEw lillnJssa33ns sEM lua!led aqd 'sglaJ3E srnoq 8p pa$3a*apQE31ug3 ‘paalq padelap E SEM aposrda alaaas Jaq$o ayJ, 'papE3 (IuaJrnD uogElnSEo3 rEIodouour PUB ‘ddelaql dexg ‘uo!$Da[u!OOO‘~JI:Iaugdauida ‘apEuodwE$ uooIIEq) lidexayl 3ldoxopua paqnbal
aSEy~~owaq
-Ed JO %v'r
Ial3E uogezgoqtua
O!yde&lo@uE
JO apostda aJaAas au0
uj paJJnxo
's)uag
ampaOoJdlsod
-EAxasuo3 paSEuEw
aJaM sggEaJwEd
ayl lal3E s&p
*sggEa~~uEd ampaDoId$sod $0~ ‘~5.0 = d)
.(sdnolEi snoaueIla+ur pue ‘Journ) ‘(s)auo~s Imp uounuo~~ pau!quxoD aql ‘SA dn0.B aOS) ~00’0 > dq ~suo~~empu~zngmds ~03examaaS,, so&/srr
(96-&l.) E'S9
(001)EZP
(98-67)E'ES (0'8) PE /98-P&)Z'S9 (Z'PZ) 09 (SG-ZZ) 8'Z9 ki'9E~ ES1 &6L-Et)L'tP (Z.6'2) 991
1WO.L
.snoauerIamg.y .JournL
(s)auWPnp~o~~o3 uowun3sLp aos
uo~~Eu1~03 67, pue &I
sluaged OM$J~$JEI 30 sapostda alaAas
‘a$ElapouI p ‘pI!tu ~1 alam aJaqL
(%p’g)sliep
.($ualajJ!p
luaJajj;rp uo
sekEp auIEs aqJu0 paurxo3lad SEM~OS
Axranbay
‘dno& aos
OZ/PZ t&/6Z oor/c9 ,PSI/ZZ
8ulpaalq ampaDoxd$sod
pue
IJJ~M s)ua;r$ed11~ *suo!snga IEmaId uourSaIqd q];[~ ampaDoJd
~Euo~~~ppEue~o3pazgE~~dsoqa~a~ ayL
z pue
@3gsgE~s 10 (%I'~)sB UaqM
JEI!UI!S~JIM palm330
SggEal3UEd
ay~u1~(sdno~8snoauE~~ax~wpue
'lourn
~0s
dq paunguo3
S~SOUiiiE!p aqJ pey
s$uagEd JO
%I'98 ‘II"JaA() '%Z'L6 u! (@I UUU s&7) pa$EAaIa SEM alnssald laJXI!qdS Ieseq aq,L '(Zgt/Lpt) paldrua$$E SEM
11 UIO~M
dIIn3ssaxns
II! sluaged JO %~'06 SBM
~~0s
‘dnoxtl (30s
II! paurro3lad aq$ %ourv
‘aUO~S~~npuour~o~pau~quro~ay~'s~dno~~(IO~a~~ u~S~~~~Ea~xxEd~~-~sod 3oh1anbaqay~~o3
~00'0 = d)
30 UogEnIEAa
x~southxp
pUE
dnokl (Ios aql II!paxnwo ‘uoy$ExIduro~ Juanbag
SggEalXIEd
JO sapos!da
~SOUr aq] SBM
S~J~JEaJXIEd
x!ayj 30 &~~pe&l
suo~~E3~~duro3 paxpur-Sz
JO addI aqL
?wlajJ!p ~wwww 10~‘cio*o < d) kh0.d SB +I0
= d)suog
0) JSEJJUOLI U! s)uagEd
‘(SaJEp!pUED IE@JnS
103 holo~a~xr!qds
uoyyasu! lua)s ~03 arnj3nls dreg!q qg~
ay$ JO O/o~'g~‘JaAaMOH *sjuagedjO oho'p U! %~.I.m330 *p a1qEL uo palElnqe$ ale &aAas aql pue
p~epUE~SUE~~(%z'g)uo!~le~~Id~o3EU~~InSa~O~dIay~I alouIouse~duIo~ola~~u!~ds~n3a~d'(600'0 -EXpUy Jay10 IIT? “03 %L’IT
aos
JO %r'rz u! paqnbal SEM $1 '(E a[qxL) SwatlEd JO %f,'sr uy paurxo3xad SCM huolola~wgds JnDaxd IOU
‘so.0 < d) (%s*II)
uo JO (%z’o~)sz ~0s
sdep w=yj!p
ac[l aaoJaq dIa~Erpauuu! paurxo3
uayM
huanbaq
sy$EaJxIEd
Iward
sggearxred
3p101q3 ‘(p)auro;rpudsdwns
‘(~)ddE~aq$
'(PC) suoseal snoaue~~aw~ur (p) pue
!(~)loU~n] hIlndure
pue ‘(Et)Eu.ro~dti-[10 euroup
(3)
ggt)aos
(e)(r aIqEJ,)a.IaMhuo~o~a~wryds
ay] ~03 suogexpu!
aqL
*slua;r)ed czf, uo paurlojlad
pue 6862
SEM SB ‘0665 wnsnv
q~wv
uaawaa
Ienba ~J!M palm3
‘dno&i (Ios ayl XII s$uagEd
pa~~nba~uogex~dwo~ $0~ ‘Z~~YJ
~03 0Qj.z HUE
ayl
ay$ JO auoN = d)
SUO~JEX~U! ~IIE~!J~!JE~S
‘(~ua~agp
%&‘E
‘(gp()‘O
= d)
SaUOJS
JO3 %6-v 01 $SEQUO3 KIT%8'()t uy paJm330
10~ ~3 lage suogE3gduro3
a[os
*(fj()o*o = d tsdnoh'snoau
‘Jotunl ‘(s)auolsIx-tp uoun1103 pautq 103 pawo3lad
JaqJO dUE
SEM ~3 uayM
suoyExlduro3
dpuanba;r3
‘JaAaMOH
Sz laJJE a)EJ uogexIduro3
~~~dno&iaos
paUIaap IOU sluayed) (P)
~~uo~y~‘(~)a~a~o~~opa~o~~‘(~)uo~~.~asu~s~saq~so~dop -Ua “03 yEaI aI!q ‘(01) IEAOUIaJ aUO$S 3gEa;rXIEd JO3 luajs ~o/pue uoge~tp ~03 s@!h1~~oq3 Su!soraps hxu -!Jd kpnpuy
-183 3!~E~SE~a~‘(OZ)EuIOU!~~~O~~UEIOq~‘(pZ)E~OU!~ -;re~~!~ea~3uEdSu!pnI~u;r ‘(09)(uo~~~asu~s~saq~so~dop !(cgT)(s)auols 13np uounuo3 (q) -ua ~03) JownJ
t(s$uayEd QIEWYEP -lad SEM
-30 suogEx~duro3 iikrou_rv *ha%ns
*(luaJajjrp QE~+~)EJS snoaue~~ax~w
IOU '6L()'O= d) SJOUInJ “03
hy~q
-ElIax;rurpue
-III03ayi “03 %E’~T 'SA %Q'O~) UOnEXpU! ~03 ueq~ aos
a.IOUIpa.Im~~O Oj6.9 SEM
'(7,aIqEL)
IIElaAO aqJ,
‘(LOO*0 > d !uauroM yo8*zfj)axreururopald aIEura3 paynxu pue
(roo*f~ > d:s~Iead~.p~‘uEaur)a8e~a~unoddq(~a~qE~) pazpalxsey3
ys~guo3u1~(~~u.1u.10~>
ayl ‘sdno&llaq$o aylo)
o~B~u1u1~~7
s~~a~nssa~d
ayl ‘s)uagEd 30 Oh8.P IEUO!$~~E UE
la$Bu!qds IWEq
.palmbar hra%ns JO q30ys xldas ‘alaaas pm !uoguaMalu! snoaue~nmad JO ~!do~sopua~odderarIlIel!dsoq30 sleepE< Bu!~nba~ssauIl! ggdas JO aplqa3 ‘almapour !ampaDold aql lal3e sleep E 5 103 3$3& < amlmadmal ‘p~pu se papeti SBM uog3a3uI *pa+nbaJ (p=?@ms JO snoauelnxad) uo!lualwalu! JO sleep 1~17 103 dde -lay4 lexpam ‘alahas pm :shp 01-t ~03 @cnpatu paleaq uogeJo3Jad algap due ‘awapotu !shp E r ~03 uogms pm spy3 dq paleaq 1seguo3 ~0 pyj 30 yz?aI @!!Is haa Quo JO alq!ssod‘P1!":SMOIIOJse papeti Sl?Muoyxo3.Iad jpz$hs JO 3yde~Bo@he) pal;mbal uo;rwahIalur JO sIIa3 pal paqxd 30 n 5 z 30 uo!sn3sueq ‘aJaaas pm !pagnbar hafhns JO Uo~~uaAJa~u~ 3ydeJSo@Ie ou ql;[~‘SIla:, poolq pal paqxd 30 n fF JO uo!sn3sueJJ ‘aJeJapouI :paJFnbaJ uo!sysue~ 0U qJ!M ‘8 E> 30 IaAaI u!qo@omaq UT aseamap e ~J!M uo!~e!~osseu;r s!sauralewaq ~0 malam se qms %npaaIq
30 amapyaa
U~!i!!HUIUI(-J~~ amssa~d~a~~u~yds~EseqEpEq~o~~~~)
(sydomopua
IOU ‘.a.!) pxy.up
t ‘ON ‘LOL ‘IOA 63070EI.L.N30XLSV3
‘p~itu se papeti
‘TV 13 mms
OLOI
October
COMPLICATIONS
1991
Table 2. Endoscopic
Sphincterotomy-Znduced
Complications Complications
Indication
No. of ES
SOD Common duct stone(s) Tumor Miscellaneous
166 163 60 34
Total
423
and 48-Hour
OF ENDOSCOPIC
SPHINCTEROTOMY
and 30-Day Mortality
Surgery
48-h mortality
39day
mortality
(%I”
(%Jb
(%)I
10.8d 4.9 3.3 2.9
0 0 0 0
0 0 0 0
0.0, 0.6 6.7 5.9
6.9
0
0
1.7
(%I
“Refers to the percentage of patients requiring surgery for the complication. bRefers to the percentage of patients dying within 48 hours of the ES. “Seven patients were lost to follow-up and were excluded from tabulation, Six of the seven deaths were seemingly sphincterotomy, whereas one was possibly related. dP = 0.009,SOD group vs. the combined common duct stone, tumor, and miscellaneous groups. "P = 0.03,SOD group vs. combined common duct stone, tumor, and miscellaneous groups.
and treatment with a Bicap hemostatic probe (CirconACMI, Stamford, CT) on two occasions (the first session with Bicap controlled the bleeding for only 2 days). Angiography failed to identify a bleeding source. Four episodes of infection (0.9% of patients) occurred. Two patients experienced cholangitis 8 and 25 days after ES and stent insertion for pancreatic carcinoma. Both patients were managed successfully with antibiotics and stent change. One SOD patient had transient low-grade fever and abnormal hepatic chemistries after a precut ES. This patient improved on IV antibiotics and was discharged from the hospital 4 days later. An 86-year-old nursing home patient died of probable pancreatic sepsis 14 days after sphincterotomy and 7 days after extraction of pancreatic duct stones. She had been clinically stable for 13 days after the ES. Autopsy was denied. Although the exact source for her terminal febrile illness was not clinically apparent, we are tallying this as an ESrelated mortality. Perforation occurred in two patients (0.5%). One patient was managed with nasogastric suction, nasobiliary tube drainage, and gut rest, whereas the other was successfully treated by nasogastric suction and
Table 3. Early Complications
1071
of Precut Sphincterotomy Complications
Precut ES
Precut
No precut
Indication
No. of ES
(%I0
(%I
(%I
SOD All othefl
166 257
21.1 11.7'
8.6 3.3d
11.5 4.4'
Total
423
15.4
6.2
7.0
NOTE. Precut sphincterotomy was performed over a previously placed pancreatic stent in 63% of precut cases. “Refers to the percentage of patients undergoing sphincterotomy who required the precut technique. ‘Combines the common duct stone, tumor, and miscellaneous groups. "P = 0.009;dP > 0.05;"P = 0.012.
unrelated
to the
gut rest alone. No episodes of basket impaction or acute cholecystitis occurred in this series. Complications after ES for all indications occurred in 11 .O% of patients with a normal-diameter common bile duct (< 10 mm) in contrast to only 2.8% of patients with a dilated biliary tree (P < 0.001) (Table 5). Similarly, more SOD patients had a complication when ES was performed in the presence of a normaldiameter common bile duct (15.1% vs. 3.3%; P = 0.020). The relationship of the common bile duct diameter and the distribution of complications after ES for all patients (combined group) and for SOD patients is shown in Figures 1 and 2, respectively. For the combined group, complications occurred in 19.6% of patients with a small-diameter common bile duct (I 5 mm) in contrast to 8.2% with an intermediate size (6-< 10 mm) (P = 0.023) and 2.8% with a dilated common bile duct (210 mm) (P < 0.001 vs. small diameter and P = 0.02 vs. intermediate size). Among the SOD patients, 37.5% with a small-diameter common bile duct had a complication in contrast to 8.5% with an intermediate size (P < 0.001) and 3.3% with a dilated common bile duct (P < 0.001 vs. smalldiameter common bile duct; P > 0.05 for intermediate size vs. dilated common bile duct). Thirty-day follow-up was obtained in 98.3% of patients (Table 2). The overall mortality rate during this period was 1.7%. The mortality rate for the SOD group was 0% (P = 0.03 for the SOD group vs. the combined common duct stone(s), tumor, and miscellaneous groups) in contrast to 0.6% (P > 0.05, not statistically different), 6.7% (P < O.OOl), and 5.9% (P = 0.002) for the common duct stone, tumor, and miscellaneous groups, respectively. Six of the seven deaths were seemingly unrelated to the ES. One patient died of arrhythmia 11 days after ES and 7 days after surgical resection of a large choledochocele. Four patients with a malignancy died 12, 14, 21, and 29 days after the ES and stent placement of metastatic cancer, urosepsis, aspiration pneumonia, and metastatic cancer, respectively. One patient died of myocar-
GASTROENTEROLOGY
1072 SHFdWiANETAL.
Vol. 101, No. 4
Table 4. Type and Grade of Complications After Endoscopic Sphincterotomy Pancreatitis
Bleed Indication
Mld
Mod
Sev
SOD Common duct stone(s) Tumor” Miscellaneous”
1 0 0 0
2 1 0 0
0 2 0 0
Total
1
3
2
NOTE. See Materials “See text for specific
and Methods indications.
for definition
Sev
Mld
Mod
Sev
Mld
Mod
Sev
7 4 0 0
4 0 0 0
2 0 0 0
0 0 0 0
1 0 2 0
0 0 0 1
0 0 0 0
1 1 0 0
0 0 0 0
11
4
2
0
3
1
0
2
0
of grade of complication
(Mld, mild; Mod, moderate;
Endoscopic sphincterotomy may be an effective therapy for patients with SOD (l-4). Despite its potential use for this disease, ES has been reported to be associated with a higher complication rate when performed for SOD than when performed for choledocholithiasis (Table 6). Among the 51 patients with SOD treated by Thatcher et al. (Z), 8 complications (15.7% of patients) occurred in 7 patients (4 patients with perforations, 2 with hemorrhage, and 2 with pancreatitis). Complications were more likely to occur in the presence of a nondilated common bile duct. Surgery was required in 3 patients with perforations, but no deaths occurred. In the series reported by Neoptolemos et al. (3), 10 complications in 8 patients occurred among the 32 attempted sphincterotomies. These complications were bleeding in 3, acute pancreatitis in 4, perforation in 2, and acute cholangitis in 1 patient. Three patients required surgery, but no deaths occurred. Similarly, Leese et al. (5) reported that complications were more common when ES was performed for papillary stenosis (6 of 37 patients, 16.2%) than for common bile duct calculi (33 of 319, 10.3%). In contrast, Roberts-Thomson and Toouli (4) reported only 2 (4%) complications (bleeding and acute pancreatitis) in 50 patients treated for SOD.
Complications: Relation to Bile Duct Diameter Frequency
of CBD
Our entire
series
Sev, severe).
Geenen et al. also found the complication rate for patients with SOD (5 of 80 patients, 6.3%) to be similar to that of patients with common bile duct stones (18). In Gregg’s (19) series of 120 patients with SOD, 13 cases of pancreatitis (2 hemorrhagic), 2 episodes of bleeding, and 1 perforation occurred after ES for SOD. It has been postulated that the hypertrophic or fibrotic sphincter in patients with SOD increases the risk of pancreatitis, because extended periods of high intensity current may be required to sever the sphincter. Duodenal perforation has been reported to occur more frequently in SOD patients (19,20). Among the four perforations reported by Thatcher et al., three had a nondilated common bile duct (2). The increased risk of perforation may be a reflection of the relatively small duct size lowering the safety margin in performing a complete ES (9). Classen found a mortality of 2.2% among 813 patients undergoing ES for suspected SOD in contrast to 1% mortality among 7585 patients undergoing ES for stone disease (8). Most studies reporting the complication rate of ES for SOD are limited by their retrospective nature and use heterogenous populations and a small number of patients. Our study was undertaken to prospectively evaluate the ES-induced complication rate in a large group of patients with manometrically documented SOD in the majority (86.1% had abnormal basal sphincter pressures) and to compare it with the complication rate for other indications. The complication rate after ES was significantly higher for patients
Discussion
Diameter
Perforation
Mod
Mld
dial infarction 5 days after successful removal of common duct stones. As noted, one patient died of probable pancreatic sepsis possibly related to the sphincterotomy performed 14 days before.
Table 5. Endoscopic Sphincterotomy-Induced
Infection
Biliary
of complications
stones
SOD
Tumor
Miscellaneous
23/210 (11.0%)” 6/213 (2.8%)b
5/48 (10.4%) 3/115 (2.6%)d
16/106 (15.1%)” 2/60 (3.3%)’
2/38 (5.3%)g o/22 (O.O%)h
O/18 (0.0%)’
Total
29/423
8/163 (4.9%)
18/166
2/60 (3.3%)
l/34 12.9%)
(6.9%)
NOTE. Numbers represent the frequency of complications defined as dilated. ’ VS. bP < 0.001; ’ VS. dP = 0.035; ’vs. rP = 0.02; gvs. * and
for the given
range
’vs. ‘not statistically
(10.8%)
of common different,
bile duct
P > 0.05.
(CBD) diameter;
l/16(6.3%)'
CBD 2 10 mm was
October 1991
COMPLICATIONS
100
0
86
UI E 0
All
patients
? ?Complicstlons
90
I
I
i
70 60 1
% P
50
P
40 1
30
-I
20 10 0
1
<5
6-7
0-9
Common
10-11
Bile Duct Diameter
12-13
>
14
(mm)
Figure 1. Frequency of complications after ES for all indications: relation to common bile duct diameter [a, small (15 mm); b, intermediate (6- < 10 mm); c, dilated (2 10 mm)]. 0, All patients; ?,? complications. Percentages represent the frequency of complications for the given common bile duct diameter. a vs. b, P = 0.023;a vs.c, P < 0.001;b vs. c, P = 0.020.
with SOD than for other indications (10.8% vs. 4.3% for all other indications combined). In the SOD group, pancreatitis occurred in 7.8%, hemorrhage in 1.8%, perforation in 0.6% and infection in 0.6% of patients. In contrast, in the combined common duct stone, tumor, and miscellaneous group, pancreatitis occurred in 1 .S%, bleeding in 1.2%, perforation in 0.4%, and infection in 1.2% of patients. Our impression is that the common bile duct is more difficult to deeply and selectively cannulate in SOD patients. This may result in repeated cannulation attempts and pancreatic duct injections, which have been shown to result in a higher frequency of postprocedure pancreatitis (6,21,22). Papillary edema and sphincter of Oddi spasm resulting from mechani48
40
P
24
P 16
25
6-7 Common
8-9
10-11
Bite Duct Diameter
12-13
2
14
(mm)
Figure 2. Frequency of complications after ES for SOD: relation to common bile duct diameter [a,small (I 5 mm); b, intermediate (6- < 10 mm): c, dilated (2 10 mm)]. Cl, All patients; ? ? complications. Percentages represent the frequency of complications for the given common bile duct diameter. a vs. b, P < 0.001; a vs. c, P < 0.001;b vs. c, not statistically different (P < 0.05).
OF ENDOSCOPIC SPHINCTEROTOMY
1073
cal trauma may restrict the flow of pancreatic juice and result in pancreatic edema and pancreatitis (23). In patients with common duct stones, cannulation is often easier and pancreatography frequently not required, thus limiting mechanical trauma to the papilla. The significantly higher need for the precut technique in patients with SOD in this series supports the notion that selective cannulation of the bile duct is often difficult in these patients. The incidence of pancreatitis after SOM may be as high as 20% (14,24) and, theoretically, might contribute to the risk of this complication if performed immediately before the ES. However, when SOM was performed before ES on the same day or on different days (or not at all), the incidence of all complications and of pancreatitis was no different. However, our routine use of the SOM aspirating technique almost certainly decreased the frequency of SOM-associated pancreatitis (14). The overall complication rate in the present series was 6.9%. No patient required surgical intervention, and no immediate mortality occurred. These rates for very serious complications are somewhat lower than those reported in the literature (22,25-27). The current series seemed to have fewer serious complications compared with what was informally recalled from our initial lo-year experience in which a few cases of post-ES phlegmon, abscess, pseudocyst, or bleeding resulted in death or laparotomy. In part, our more recent complication rates may be related to greater operator experience, more attention to cannulation/ES details, and our younger (mean age, 55.3 years), relatively more fit population. A variety of precautionary measures are used in an effort to avoid complications. We strive to always assure drainage of an obstructed duct, because the risk for cholangitis is substantially higher if the duct is not decompressed (28). In the present series, nine patients had biliary stents placed because of failure to clear the common duct of stones post-ES (stones were either large or proximal to a narrowing). In five patients (8.3%) with a malignancy, drainage took place percutaneously after unsuccessful placement of stents endoscopically. These interventions almost certainly reduced the incidence of postprocedure cholangitis. Moreover, using alternative methods to fragment common duct stones (e.g., laser and mechanical lithotripsy) and placing long-term biliary stents when the common duct is not cleared of stones makes large sphincterotomies (with the associated increased risk of complications) often unnecessary. Prophylactic antibiotic therapy is continued after the procedure if an obstructed pancreatic or biliary tree is found at ERCP (29). We attempt to reduce the risk of hemorrhage by correcting a coagulopathy, discontinuing (ideally) aspirin and nonsteroidal antiinflammatory use 7-10 days before the ES, limiting the length of the
1074
SHERMAN ET AL.
GASTROENTEROLOGY Vol. 101. No. 4
Table 6. Reported Complication Rates Associated With Endoscopic
Sphincterotomy
for Sphincter of Oddi Dysfunction
No. of complications Study Thatcher et al. (2) Roberts-Thomson et al. (4) Neoptolemos et al. (3) Leese et al. (5) Geenen et al. (18)
n 51 50 32 37
Gwg(l9)
80 120
Total
370
Total
Bleed
8" 2 10b 6
2 1 3 _c
5 16
2
Pant
Perf
Chol
No. of surgeries
No. of deaths
2 1 4
4 0 2
0 0 1 -
3 0 3 -
0 0 0 -
13
0
-
1
0 -
-
47(12.70/b)
Pant, pancreatitis; Perf, perforation; Chol, cholangitis. “Eight complications occurred in seven patients. *Ten complications occurred in eight patients. “Data unavailable.
cut to the amount required, and using coagulation current (alternating with blended current) for the cephalad half of the incision or any recut endoscopic sphincterotomy. If bleeding is significant, we attempt to control it endoscopically. If endoscopic therapy fails to control the hemorrhage, angiographic embolization is quickly attempted (30). We attempt to reduce the risk of pancreatitis by limiting cannulation time; avoiding the number of injections, acinarization, and coagulation current near the pancreatic duct orifice; meticulously disinfecting endoscopes; using low osmolarity agents in patients with a prior episode of pancreatitis (or prior episode of ERCP-associated pancreatitis); premeditating with oral steroids and antihistamines in patients with previous contrast media reactions; and using the aspirating SOM catheter when SOM of the pancreatic duct sphincter is performed (22). When the endoscopist is having difficulty achieving deep biliary cannulation, a choice must be made whether to continue cannulation attempts, terminate the procedure, perform precut sphincterotomy, use percutaneous assistance, or consider surgery. Multiple factors must be considered in this choice. We tend to use precut techniques fairly aggressively. Precut ES was used in 15.4% of patients in this series. This technique has been reported to be potentially hazardous, particularly when performed by inexperienced endoscopists (3 1). However, our complication rate for precut ES was not statistically different from standard ES and is in agreement with the complication rate reported by Dowsett et al. (10) for this technique. Actually, a trend toward greater safety was observed from precutting. This may, in part, be related to the pancreatic stent placed before precutting or to the fact that additional prolonged sphincter probing was often avoided. Endoscopic sphincterotomy-induced complications occurred in 11.0% of patients with a normaldiameter common bile duct and was significantly
reduced in the presence of a dilated common bile duct. These results are in support of those of Mustard et al. (2O), Thatcher et al. (2) and Gregg (19). Patients with small-diameter (I 5 mm) common bile ducts had significantly more complications than patients with intermediate-size (6 to < 10 mm) or dilated (2 10 mm) common bile ducts. Moreover, patients with intermediate-size ducts had more adverse events after ES than patients with a dilated common bile duct. Although this separation of normal-sized common bile ducts into small and intermediate is somewhat arbitrary, it serves to emphasize the significant rate of complications for SOD patients with a small-diameter duct (37.5%) and the stepwise reduction in adverse events for those with an intermediate size and a dilated common bile duct, respectively. Thirty-day mortality was 1.7%. However, the procedure-related mortality was believed to be 0.2%. Again, these procedure-induced fatalities are lower than many reports in the literature and may be related to our younger, lower-risk population and attention to details of technique. Our data again support the impression that ES for SOD is more hazardous than for other indications. Because of the relatively high incidence of major complications, some authorities recommend a trial of medical therapy before attempts at endoscopic treatment (19). ES for SOD has been reported to result in pain relief in 50%-90% of patients (l-4). Because surgical sphincteroplasty with septoplasty severs the pancreatic and biliary sphincters, would surgical therapy for these patients be preferable? A prospective randomized study addressing the therapeutic benefits and complication rates of endoscopic vs. surgical therapy is in progress at our institution. This study will help detect the frequency of restenosis of the sphincterotomy site, a factor not addressed in this manuscript. In summary, ES for SOD is associated with an increased risk of complications and more frequently
October
1991
requires the precut technique. Compared with standard ES, precut ES (usually with stenting) was not associated with an increased frequency of complications. The risks of complications after ES appears to be less for those with an intermediate or dilated common bile duct. More importantly, the risk of a complication is considerable for a small-diameter common bile duct, particularly when ES is performed for SOD.
References 1. Geenen
JE, Hogan WJ, Dodds WJ, Toouli J, Venu RP. The efficacy of endoscopic sphincterotomy after cholecystectomy in patients with sphincter of Oddi dysfunction. N Engl J Med 1989;320:82-87. 2. Thatcher BS, Sivak MV, Tedesco FJ, Vennes JA, Hutton SW, Achkar EA. Endoscopic sphincterotomy for suspected dysfunction of the sphincter of Oddi. Gastrointest Endosc 1987;33:9195. 3. Neoptolemos JP, Bailey IS, &r-Locke DL. Sphincter of Oddi dysfunction: results of treatment by endoscopic sphincterotomy. Br J Surg 1988;75:454459. IC, Toouli J. Is endoscopic sphincterotomy 4. Roberts-Thomson for disabling biliary-type pain after cholecystectomy effective? Gastrointest Endosc 1985;31:370-373. 5. Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their management following endoscopic sphincterotomy: results in 394 consecutive patients from a single centre. Br J Surg 1985;72:215-219. 6. Podolsky I, Haber GB, Kortan P, Gray R. Risk factors for pancreatitis following ERCP: a prospective study (abstr). Am J Gastroenterol 1987;82:97A. 7. Siegel JH. Endoscopic papillotomy in the treatment of biliary tract disease: 258 procedures and results. Dig Dis Sci 1981;26: 1057-1064. 8. Classen M. Endoscopic papillotomy-new indications, shortand long-term results. Clin Gastroenterol 1986;15:457-469. 9. Krims PE, Cotton PE. Papillotomy and functional disorders of the sphincter of Oddi. Endoscopy 1988;20:203-206. 10. Dowsett JF, Polydorou AA, Vaira D, D’Anna LM, Ashraf M, Croker J, Salmon PR, Russell RLG, Hatfield ARW. Needle knife papillotomy: how safe and how effective? Gut 1990;31:905908. 11. Dowsett JF, Vaira D, Hatfield ARW, Cairns SR, Polydorou A, Frost R, Croker J, Cotton PB, Russell RLG, Mason RR. Endoscopic biliary therapy using the combined percutaneous and endoscopic techniques. Gastroenterology 1989;96:1180-1186. 12. Geenen JE, Hogan WJ, Dodds WJ, Stewart ET, Arndorfer RC. Intraluminal pressure recording from the human sphincter of Oddi. Gastroenterology 1980;78:317-324. 13. Sherman S, Troiano FP, Hawes RH, Lehman GA. Does continuous aspiration from an end and side port in a sphincter of Oddi manometry catheter alter recorded pressures? Gastrointest Endosc 1990;36:500-503. 14. Sherman S, Troiano FP, Hawes RH, Lehman GA. Sphincter of Oddi manometry: decreased risk of clinical pancreatitis with use of a modified aspirating catheter. Gastrointest Endosc 1990;36:462-466. 15. Carr-Locke DL, Gregg JA. Endoscopic manometry of pancreatic
COMPLICATIONS
OF ENDOSCOPIC SPHINCTEROTOMY
1075
and biliary sphincter zones in man: basal results in healthy volunteers. Dig Dis Sci 1981;26:7-15. 16. Guelrud M, Mendoza S, Rossiter G, Villegas MI. Sphincter of Oddi manometry in healthy volunteers. Dig Dis Sci 1990:35:3840. comparisons and confusion: com17. Cotton PB. Complications, mentary. In: Cotton PB, Tygat GNJ, Williams CB, eds. Annual of gastrointestinal endoscopy, 1990. London: Current Science Ltd., 1990:7-g. 18. Geenen JE, Hogan WJ, Dodds WJ. Sphincter of Oddi. In: Sivak MV, ed. Gastroenterologic endoscopy. Philadelphia: Saunders, 1987:735-751. 19. Gregg JA. Function and dysfunction of the sphincter of Oddi. In: Jacobson IM, ed. ERCP: diagnostic and therapeutic applications. New York: Elsevier Science, 1989:139-170. 20. Mustard R, Mackenzie R, Jamieson C, Haber GB. Surgical complications of endoscopic sphincterotomy. Can J Surg 1984; 27:215-217. 21. Hamilton I, Lintott DJ, Rothwell J, Axon ATR. Acute pancreatitis following endoscopic retrograde cholangiopancreatography. CIin Radio1 1983;34:543-546. 22. Sherman S, Lehman GA, ERCP- and endoscopic sphincterotomy-induced pancreatitis. Pancreas 1991;6:350-367. Nordenstam CG, Brackett 23. Saari A, Kivisaari L, Standertskjold, K, Schroder T. Experimental pancreatography: a comparison of three contrast media. Stand J Gastroenterol 1988;23:53-58. 24. Rolny P, Anderberg B, Ihse I, Lindstrom E, Olaison G, Arvill A. Pancreatitis after sphincter of Oddi manometry. Gut 1990;31: 821-824. 25. Tedesco FJ, Vennes JA, Dreyer M. Endoscopic sphincterotomy: the USA experience. In: Okabe H, Honda T, Oshiba F, eds. Endoscopic surgery. New York: Elsevier Science, 1984:41-46. 26. Vaira D, D’Anna L, Ainley C, Dowsett J, Williams S, Baillie J, Cairns S, Croker J, Salmon P, Cotton P. Russell C, Hatfield A. Endoscopic sphincterotomy in 1000 consecutive patients. Lancet 1989;2:431433. 27. Geenen JE, Vennes JA, Silvis SE. Resume of a seminar on endoscopic sphincterotomy (ERS) Gastrointest Endosc 1981;27: 31-38. 28. Ostroff JW, Shapiro HA. Complications of endoscopic sphincterotomy. In: Jacobson IM, ed. ERCP: diagnostic and therapeutic applications. New York: Elsevier Science, 1989:61-73. 29. Hershey SD, Sugawa C, Cushing R, Ledgerwood AM, Lucas CE. The value of prophylactic antibiotic therapy during endoscopic retrograde cholangiopancreatography. Surg Gynecol Obstet 1982;155:801-803. 30. Saeed M, Kadir S, Kaufman SL, Murray RR, Milligan F, Cotton PB. Bleeding following endoscopic sphincterotomy: angiographic management by transcatheter embolization. Gastrointest Endosc 1989;35:300-303. risky technique for experts 31. Cotton PB. Precut papillotomy-a only. Gastrointest Endosc 1989;35:578-579.
Received October 24,199O. Accepted February 4,199l. Address requests for reprints to: Glen A. Lehman, M.D., Indiana University Hospital, 926 West Michigan Street, Room N541, Indianapolis, Indiana 46202. Dr. Stuart Sherman was the recipient of an American Society for Gastrointestinal Endoscopy/Olympus Corporation of America Tier III scholarship.