intravenous morphine for augmentation of postoperative t-tube cholangiograms in liver transplant...

Intravenous Morphine for Augmentationof Postoperative T-tube Cholangiogramsin Liver Transplant Recipients withCholedocho-choledochal AnastomosesWael E.A. Saad, MBBCh, Ian J. Wilson, MD, Mark G. Davies, MD, PhD, Karin E. Westesson, MD,

Wael M. Darwish, MBBCh, and David L. Waldman, MD, PhD

PURPOSE: To determine the effectiveness of augmenting T-tube cholangiography by using intravenous morphine inorthotopic liver transplant recipients with choledocho-choledochostomies and poor filling of intrahepatic biliary ductsand to determine factors that may increase the likelihood of nonfilling of intrahepatic ducts.

MATERIALS AND METHODS: A retrospective review of T-tube cholangiograms obtained in orthotopic livertransplant recipients was performed. Intravenous morphine had been given by two of five operators to augmentT-tube cholangiograms with poor filling of bile ducts. Patients with malpositioned tubes and decompressive bileleaks were excluded from morphine diagnostic efficacy evaluation but were included in the overall cholangiogramdiagnostic yield. Anastomotic narrowing, if present, was graded as follows: >50%, 20%–50%, and <20% diameterreduction. Patients with intrahepatic bile duct filling were compared to those without filling with regard to age, sex,time from transplantation, and clinically significant (>50%) stenoses.

RESULTS: One hundred sixty-eight cholangiograms were obtained in 127 recipients. Twenty-three of the 168cholangiograms (13.7%) had malpositioned/blocked T-tubes and five (3%) had decompressive leaks; 140 cholangio-grams had well-positioned tubes and no leaks. Twenty-two of the 140 cholangiograms with well-positioned tubes andno leaks (15.7%) had nonfilling of peripheral bile ducts. Morphine (range, 2–6 mg; mean, 4 mg) had been used in 13cases. Adequate filling after morphine was noted in 12 of the 13 cases (92%), and no complications occurred. Morphineimproved adequate diagnostic examination of well-positioned patent T-tubes from 85% (123/145) to 93% (135/145). Noparameters helped predict inadequate filling in well-positioned tubes (P > .05).

CONCLUSIONS: In 92% of cases, intravenous morphine was successful in opacifying the biliary tract withoutcomplications. In well-positioned T-tubes, the use of morphine increased diagnostic yield from 85% to 93%. Nopredictors for inadequate filling were found.

J Vasc Interv Radiol 2009; 20:1320–1328

POSTOPERATIVE T-tube cholangiog-raphy after liver transplantation isused to evaluate the biliary tract non-

From the Department of Imaging Sciences, Univer-sity of Rochester Medical Center, 601 Elmwood Ave,Box 648, Rochester, NY 14642 (W.E.A.S., I.J.W.,W.M.D., D.L.W.); the Department of CardiovascularSurgery, Methodist DeBakey Heart and VascularCenter, The Methodist Hospital, Houston, Texas(M.G.D.); and the University of Rochester School ofMedicine and Dentistry, Rochester, New York
(K.E.W.). Received November 10, 2008; final revi-sion received June 24, 2009; accepted July 13, 2009.
1320

invasively in search of operative com-plications in the postoperative periodand/or to divert bile from the anasto-

Address correspondence to W.E.A.S.; E-mail:[email protected]

M.G.D. has a speaker and general consultant agree-ment with Boston Scientific. None of the other au-thors have identified a conflict of interest.

© SIR, 2009

DOI: 10.1016/j.jvir.2009.07.025

mosis to promote healing (1–3). Thesesurgically placed small-caliber (usu-ally �6 F) tubes can be placed in var-ious ways and configurations (1–3).When placed for diagnostic purposes,they act as a portal for contrast me-dium injection (2,3). This is importantin liver transplant recipients in thepostoperative period because it is notuncommon to have nonspecific in-creased liver function tests. A low-risk, noninvasive cholangiogram canrule out biliary disease as the cause ofliver dysfunction. Furthermore, evalu-
ating the intrahepatic bile ducts can
mailto:[email protected]

tom

Saad et al • 1321Volume 20 Number 10

show diffuse biliary disease that re-flects a global graft complication suchas ischemia or reperfusion injury. Oc-casionally, radiologists are unable toobtain a diagnostic cholangiogram af-ter surgery. This may be due to (a)T-tube blockage, (b) T-tube dislodge-ment/malposition, and (c) leakage ofbile out of the biliary tract through aninjury, from a leaking T-tube, or ana-tomically through the sphincter ofOddi.

Morphine has been used in biliaryscintigraphy (nuclear hepatobiliaryscanning) to contract the sphincter ofOddi to “backfill” the biliary tract withradiotracer to visualize the gallbladderat these examinations (4–6). This is tohelp prove cystic duct obstruction,which aids in the diagnosis of chole-cystitis (4–6).

We performed this study to evalu-ate the efficacy of converting these po-tentially nondiagnostic posttransplantcholangiograms to diagnostic cholan-giograms by administering intravenousmorphine, which contracts the sphincterof Oddi and may allow better opacifica-tion of the entire biliary tree due to back-filling. In addition, we evaluated thefrequency of nondiagnostic T-tube chol-angiograms due to biliary escape intothe small bowel in postoperative livertransplant recipients with choledocho-choledochal anastomoses and an intactsphincter of Oddi and attempted toidentify factors that may increase thelikelihood of obtaining these nondiag-nostic T-tube examinations.

MATERIALS AND METHODS

Study Design

This is a retrospective study per-formed by auditing a prospectivelymaintained database of radiologic im-ages, charts, and records of adult livertransplant recipients with choledocho-choledochal anastomoses (recipientsphincter of Oddi in place) who un-derwent postoperative T-tube cholan-giography from July 2005 to June 2008(3 years). Institutional internal reviewapproval was obtained, categorizingthe study as secondary use of pre-ex-isting anonymous data.

The age and sex of the liver trans-plant recipients and the time betweenliver transplantation and T-tube cholan-giography were noted. Patients were
classified into three groups: (a) recipi-
ents with dislodged and/or occludedT-tubes, (b) recipients with adequatelyplaced and patent T-tubes with a de-compressive anastomotic biliary leakthat does not allow filling of the intra-hepatic ducts, and (c) recipients with ad-equately placed and patent T-tubeswithout biliary leaks.

Transplant recipients with anasto-motic bile leaks were considered to haveadequate diagnostic cholangiogramsand were included in the overall diag-nostic yield of T-tube cholangiograms.However, they were excluded (alongwith the patients with malpostioned/blocked T-tube cholangiograms) in theevaluation of the efficacy of morphine tovisualize/opacify the biliary tract. Thisis because intravenous morphine wasnever used and its effect cannot be as-sessed in the presence of a decompres-sive biliary leak.

Surgical Anatomy

All patients were adult cadavericorthotopic liver transplant recipientswith end-to-end choledocho-choledochal

Figure 1. Fluoroscopic image from T-tuberecipient. The T-tube (seen coiled outside thtract heading distally toward the bowel viaThe cystic duct stump communicates withunder the end-to-end choledocho-choledoanastomosis is the donor or graft commonis a long stump hepatico-choledochal anas

anastmoses (Fig 1). The cystic stumps of

both the graft (donor) and recipient (na-tive) were usually not excised. One wasoversewn and the other was the portalof entry of the surgically placed Turcottube (Fig 1). The Turcot tube is not aT-shaped tube but is referred to by thisstudy as a T-tube. The tube is almostalways placed in the central bile ductsheaded toward the small bowel but notpassing the sphincter of Oddi.

T-Tube Cholangiography andDefinitions

All recipients received prophylacticintravenous antibiotics. Patients re-ceived 3.375 g of intravenous Zosyn(Wyeth Pharmaceuticals, Philadelphia,Pennsylvania; pipercillin and Tazobac-tam), and, if allergic to penicillin and/orcephalosporins, 400 mg of intravenousciprofloxacin was given. Subsequently,contrast medium (Conray-60; Mallinck-rodt, Tyco Health, St Louis, Missouri)was instilled via gravity drip into theindwelling T-tube. Cholangiograms areobtained at different obliquities. Turn-ing the patient and tilting the angiogra-

olangiography in an adult liver transplantatient’s body, arrowhead) enters the biliaryrecipient’s cystic duct stump (solid arrow).recipient common bile duct (R. CBD) just

al anastomosis (open arrow). Above theatic duct (G. CHD). Strictly speaking, thisosis.

che pthethechhep

phy table are rarely performed at our

e in ageed

1322 • IV Morphine for Post-liver Transplant T-tube Cholangiograms October 2009 JVIR

institution. An adequate image was de-fined as one in which the recipient anddonor common bile ducts/common he-patic ducts (with the anastomosis in be-tween), the confluence of the right andleft main bile ducts, and the right-sidedintrahepatic (anterior and posterior seg-ments) peripheral bile ducts were visu-alized. Nonvisualization of the left sec-ond-order and peripheral bile ducts wasconsidered not necessary for an ade-quate cholangiogram but was necessaryto define a complete T-tube cholangio-graphic examination (Fig 2).

Anecdotally, not seeing peripheralleft-sided bile ducts is not an infre-quent finding (see results below) (Fig2). In addition, our definition of diag-nostic adequacy is the visualizationand adequate assessment of the surgi-cal anastomosis (for patency and/orleaks) and having a good representa-tive visualization of the peripheral bileducts to rule out diffuse intrahepaticbiliary disease/liver (hepatic graft)

Figure 2. (a) Fluoroscopic image from T-trecipient common bile duct (CBD) only anopacified. The contrast medium decompresimage from T-tube cholangiography in ansphincter of Oddi has contracted and trunccontrast medium now opacifies both the rstenosis is seen at the anastomosis (betweleft-sided ducts are not seen. However, thsolitary intrahepatic stricture is seen (dash

disease and not necessarily visualizing

the entire intrahepatic biliary tree. Onthe basis of the above, the degree ofcontrast medium filling (opacificationof the biliary tree) was assessed by theoperator at three separate anatomicstations: (a) the common hepatic/com-mon bile duct including the surgicalanastomosis (central extrahepatic bili-ary stalk), (b) the confluence of theright and left main bile ducts, and (c)the intrahepatic bile ducts (right vs leftor both). On nonvisualization of theperipheral bile ducts, two of the fiveoperators administered intravenousmorphine to contract the sphincter ofOddi and backfill contrast mediumthat was escaping through the sphinc-ter and into the bowel. Two to fourmilligrams of morphine were injected,and this was repeated as needed. Con-trast medium continued to be instilledvia gravity after morphine administra-tion.

Biliary colic and excess sedationwere assessed after morphine admin-

e cholangiography in an adult liver transpreaches the end-to-end biliary anastomosianatomically through the sphincter of Oddilt liver transplant recipient after intraven

s abruptly (arrowhead). Contrast mediumpient common bile duct (CBD) and the gsolid arrows). The left main bile duct is strahepatic bile ducts are seen and this im

arrow).

istration. Postcholangiography com-

plications were noted. These compli-cations included cholangitis (fever�38.5°C), sepsis (hypotension requir-ing admission and fluid resuscitation),and bleeding.

The surgical anastomosis was as-sessed, if visualized, for narrowing.Anastomotic stensosis/patency wasgraded by the operator as more than50% (significant stenosis), 20%–50%,and less than 20% diameter reductionof the narrowest of the donor or recip-ient common bile ducts.

The sub-cohort with adequatelyplaced patent T-tubes was classifiedinto those with filling of the peripheral(intrahepatic) bile ducts (adequatecholangiogram) and those without fill-ing of the intrahepatic ducts (inade-quate examinations). These two groupswere compared for age, sex, time fromtransplantation, and presence of sig-nificant stenoses by using the t testwith Welch connection (for days fromtransplantation and age) and the �2

t recipient. Contrast medium opacifies thenly. None of the graft biliary anatomy isd into the small bowel (B). (b) Fluoroscopicadministration of 4 mg of morphine. The

w backfills in the biliary tract. As a result,biliary anatomy. A moderate (20%–50%)at its base (open arrow), and peripheralis considered diagnostically adequate. A

ub land s o

ses anadu ousate noeci raften een

test (for stenoses and sex). This com-

ine


parison was done to evaluate for fac-tors that would increase the likelihoodof obtaining an inadequate cholangio-gram (inadequate filling) despite havingan adequately placed patent T-tube.

The group without filling was eval-uated for those recipients who under-went intravenous morphine adminis-tration and those who did not. Contrastfilling at T-tube cholangiography wascompared before and after morphineadministration to asses for the successof achieving a diagnostically adequateT-tube cholangiogram. A comparisonof the sex, age, time of examination,and presence of significant stensoseswas performed within the group with-out filling to compare those who re-ceived morphine and those who didnot by using the t test with Welchconnection (for days from transplanta-tion and age) and the �2 test (for ste-noses and sex). This comparison wasdone to evaluate for potential selectionbias. This was done despite the factthat the decision as to which patientreceived morphine and which patientdid not was not based on patient char-acteristics or surgical consequencesbut was strictly operator-dependant.Two operators (W.E.A.S., I.J.W.) usedmorphine in an attempt to obtain anadequately diagnostic cholangiogram,and the remaining three operatorsdeemed nonfilling as nondiagnosticand terminated the examination.

RESULTS

One hundred sixty-eight T-tubecholangiograms were obtained in 127adult liver transplant recipients withcholedocho-choledochal anastomoses(mean, 1.3 T-tube cholangiograms perrecipient; range, 1–-3 cholangiogramexaminations per recipient). Eighty-five recipients had one cholangio-gram, 34 had two cholangiograms,and five had three cholangiograms.The overall population included 38women (30%) and 89 men (70%). Themedian patient age was 55 years(range, 19–72 years; standard devia-tion, 12 years). The median time fromtransplantation was 78 days (range,1–113 days; standard deviation, 30days).

Twenty-three of the 168 T-tube chol-angiograms (13.7%) showed blocked(n � 11) or malpositioned (n � 12) T-tubes. Nine patients had either blocked
(n �2 ) or malpositioned (n � 7) T-tubes
on their first T-tube cholangiogram, and14 additional patients had either blocked(n � 9) or malpositioned (n � 5) T-tubeson their second T-tube cholangiogram.Five patients (five T-tube cholangio-grams in five patients), representing 3%of cholangiograms (5/168) and 4% oftransplant recipients, had biliary anasto-motic leaks.

In summary, 28 of the 168 T-tubecholangiograms (16.7%) showed mal-positioned T-tubes (n � 12), blockedT-tubes (n � 11), or decompressivebile leaks (n � 5). These three sub-groups were excluded from the intra-venous morphine efficacy evaluation.As a result, this leaves 140 T-tube chol-angiograms (140/168, 83.3%) obtainedin 113 patients with adequately posi-tioned patent T-tubes and without de-compressive bile leaks.

Twenty-two of the 140 T-tube chol-angiograms with patent, well-posi-tioned T-tubes (15.7%) had nonfillingof peripheral intrahepatic bile ducts(inadequate T-tube cholangiogram),and 118 (84.3%) were adequate T-tube

Table 1Comparison between T-tube CholangiogIntrahepatic Bile Ducts and Those withpositioned Patent T-tubes and without D

Parameter

AdIntrahe

Duct Fill

Age (y)MedianRange 1SD

SexF 28M 63

Time between transplantationand cholangiography (d)

MedianRange 1SD

Anastomotic stenosis diameterreduction

�50% 820%–50% 37�20% 73

Note.—Data are for 140 T-tube cholangioNumbers in parentheses are percentages.percent reduction in diameter of the sma(graft) central bile ducts (common hepati* Not statistically significant (P �.05); obtconnection.† Not statistially significant (P �.05); obta

cholangiograms. The comparison be-

tween these two subgroups is shownin Table 1. Of the 22 T-tube cholangio-grams (22 cholangiograms in 22 trans-plant recipients) with nonfilling of theperipheral bile ducts, 13 were ob-tained with intravenous morphine andnine were not. Morphine was adminis-tered strictly according to operator pref-erence. Two operators always usedmorphine during the study period (13cases using morphine performed bytwo operators); none of the remainingthree operators used morphine in thestudy period (nine cases not usingmorphine performed by three remain-ing operators). No transplant recipienthad an inadequate T-tube cholangio-gram twice. This was probably be-cause inadequate T-tube cholangiog-raphy prompted T-tube removal bythe transplant service (no clinical util-ity for maintaining it). No recipientreceived morphine at two separate T-tube cholangiographic examinations.The comparison between those who wereadministered morphine and those whowere not is shown in Table 2.

s with Adequate Filling of thedequate Filling in Patients with Well-ompressive Bile Leaks

ateic Bile(n � 91)

InadequateIntrahepatic Bile

Duct Filling (n � 22) P Value

54 .129*2 19–70

14

1) 6 (27) .058†9) 16 (73)

84 .163*3 14–96

27

) 2 (9) .182†) 11 (50)) 9 (40)

ms in 113 liver transplant recipients.� standard deviation. Percentages are

t of the recipient (native) or donoruct and/or common bile duct).ed with the t test with Welch

d with the �2 test for independence.

ramInaec

equpat

ing

569–711

(3(6

75–1130

(7(31(62

graSD

llesc dain

Of the 22 T-tube cholangiograms

r i


classified as inadequate, seven hadcontrast filling to the choledocho-cho-ledochal anastomosis, 11 had contrastfilling the common hepatic bile duct

Table 2Comparison between Recipients who Rein Patients with T-tube CholangiogramsIntrahepatic Bile Ducts with Well-positiDecompressive Bile Leaks

Parameter

InadequatBile Duc

Morphi

Age (y)MedianRange 1SD

SexF 5M 8

Time between transplantationand cholangiography (d)

MedianRange 1SD

Anastomotic stenosisdiameter reduction

�50%20%–50%�20%

Note.—Data are for 22 T-tube cholangiogNumbers in parentheses are percentages.* P value was obtained with the t test wit† P value was obtained with the �2 test fo

Table 3Highest Location of Contrast Medium Fwith Inadequate Filling of the Biliary T

Cases of InadequateCholangiogram

Highesm

At SurgicaAnastomos

Morphine success whengiven (n � 12)

7*

Morphine failure whengiven (n � 1)

0

No morphine given (n � 9) 0

Note.—Data are for the 22 patients who itract. Adequate filling of the biliary tractcommon bile duct, graft common hepaticducts (peripheral bile ducts).* One of these seven T-tube cholangiogra(�50% diameter reduction) anastomotic sadministration of morphine.

without opacification of the conflu-

ence of the right and left main bileducts, and four had contrast opacify-ing the confluence of the main bileducts (Table 3). Of 118 T-tube cholan-

ved Morphine and Those who did notth Inadequate Filling of thed Patent T-tubes and without

ntrahepaticlling with(n � 13)

InadequateIntrahepatic Bile

Duct Fillingwithout Morphine

(n � 9) P Value

56 .397*8 21–70

15

8) 1 (11) .157†2) 8 (89)

79 .640*95 15–96

28

0 .012†27

s in 22 liver transplant recipients.� standard deviation.elch connection.

ndependence.

ng in the 22 T-tube Cholangiograms

ocation of Contrast Medium before-hine Administration (If Given)

At Graft CommonHepatic Duct

At the Confluenceof the Main

Hepatic Ducts

4 1

1 0

6 3

ially had inadequate filling of the biliaryefined as opacification of the recipientct, and right-sided intrahepatic bile

opacified and showed a significantosis after the intravenous

giograms that showed adequate filling

of the biliary tract, 88 (63% of all ex-aminations with well-placed patent T-tubes without morphine and 75% ofadequate examinations without mor-phine) had contrast filling of both theright- and left-sided intrahepatic pe-ripheral bile ducts. Thirty T-tube chol-angiograms (21% of all examinationswith well-placed patent T-tubes with-out morphine and 25% of adequateexaminations without morphine) hadcontrast filling of the right-sided intra-hepatic peripheral bile ducts and theleft main bile duct only (nonopacifica-tion of the intrahepatic/peripheralleft-sided bile ducts).

The 13 patients who received intra-venous morphine were given 2–6 mgof morphine, with a mean of 4 mg perliver transplant recipient. One patientreceived 2-mg of morphine and onepatient received 4-mg followed by anadditional 2-mg of morphine. The re-maining 11 patients received 4 mg ofintravenous morphine. Of the 13 pa-tients who received morphine, contrastfilling before morphine administrationwas to the level of the anastomosis inseven patients, to the common hepaticduct in five patients, and to the con-fluence of the main ducts in one pa-tient (Table 3). The intrahepatic bileducts were successfully opacified in 12of the 13 patients (92%). Ten of the 12T-tube cholangiograms rendered ade-quate due to morphine administrationhad contrast medium filling both theright- and left-sided intrahepatic pe-ripheral bile ducts (after morphine ad-ministration) (Figs 3, 4). Two of the 12T-tube cholangiograms rendered ad-equate after morphine administra-tion had contrast medium filling theright-sided intrahepatic peripheralbile ducts only (after morphine admin-istration) (Fig 2). In one of the sevencases in which contrast mediumreached the surgical anastomosis (butdid not opacify it) a significant anas-tomotic stenosis was exposed (Table3). The unsuccessful case received 4mg of intravenous morphine followedby 2 mg intravenous morphine. Nocomplications were encountered in the13 patients who had morphine admin-istration, and none had excessive se-dation from the relatively low dosesof morphine given. No patient com-plained of pain (biliary colic) afterthe morphine administration. Noneof the 13 patients had cholangitis af-

ceiwi

one

e It Fine

509–613

(3(6

854–27

292

ramSDh W

illiract

t Lorp

lis

nitis ddu

msten

ter T-tube cholangiography, although


four patients (all with adequate T-tubecholangiograms with filling of the in-trahepatic bile ducts and not requiringmorphine) who did not receive mor-phine did have cholangitis. The pa-tient with the failed morphine-aug-mented T-tube cholangiogram did not
have a history of sphincteroplasty and
was not receiving intravenous narcot-ics.

The overall diagnostic yield, whichis the rate of obtaining an adequatecholangiogram, of T-tube cholangiog-raphy after liver transplantation with-out and with the use of intravenous

Figure 3. (a) Fluorin an adult liver trafies the recipient coend-to-end biliary aanatomy is opacifietomically through tbowel (B). (b) Fluorin the same patientmg of morphine. Thgives a filling defecumn in the recipiencalled the “pseudo-the recipient commbiliary anastomosistion of the commontrast medium fills t(c) Delayed fluorosgraphic image in b4 mg of morphine.result, contrast medanastomosis (arrowbile duct and the grseen in the intrahep

morphine augmentation was 73% (123/

168) and 80% (135/168), respectively.The diagnostic yield for first time T-tube cholangiograms after liver trans-plantation was 93% (118/127). Whenconsidering cholangiograms that showpatent and well-placed T-tubes (ade-quate filling of the biliary tract) with-

opic image from T-tube cholangiographylant recipient. Contrast medium opaci-on bile duct (CBD) only and reaches the

stomosis only. None of the graft biliaryhe contrast medium decompresses ana-

sphincter of Oddi and into the smallopic image from T-tube cholangiographyin a after intravenous administration of 4phincter of Oddi has contracted andpression on the contrast medium col-

ommon bile duct (arrowhead). This isne sign.” Contrast medium still opacifiesbile duct only and reaches the end-to-endly (between arrows). Notice the disten-le duct as pressure builds just before con-graft (intrahepatic) bile ducts (cf Fig 3c).ic image from the T-tube cholangio-tained after intravenous administration of

sphincter of Oddi has contracted. As anow passes retrograde through the

nd opacifies both the recipient commonbiliary anatomy. Numerous strictures are

c bile ducts (arrowheads).

oscnspmmnad. Theoscase s

t imt cstoononbi

hecopobTheiums) aaftati

out morphine augmentation, the diag-

l-fi


nostic yield is 85% (123/145). Byadding the 12 successful cholangio-grams that were augmented by mor-phine, the overall diagnostic yield andthe diagnostic yield of patent and in-place T-tubes increase from 73% (123/168) to 80% (135/168) and from 85%(123/145) to 93% (135/145), respec-tively. In addition, when consideringsignificant anastomotic stenoses, mor-phine help detect one of 10 (10%) sig-nificant stenoses found in this study.This case of significant stenosis is oneof the seven cases in which contrastmedium only reached the anastomosisand the administration of morphineopacified the biliary tract further anddelineated the significant anastomoticstenosis.

DISCUSSION

Postoperative cholangiograms ob-tained noninvasively by injecting sur-gically placed small-bore tubes playan important role in evaluating post-operative hepatobiliary surgical com-

Figure 4. (a) Fluoroscopic image from T-trecipient common bile duct only and reachThe contrast medium decompresses anatomfrom T-tube cholangiography in an adultcholangiogram is magnified and focused o(between arrows) and the detail of the wel

plications (1,7). The existence and

known importance of these tubes areas old as hepatobiliary surgery fromthe early days of evaluating for re-tained stones in the common bile ductsafter cholecystectomy and checking onbiliary enteric anastomoses (roux en Yand Whipple procedures, etc). Theirhistory precedes that of liver trans-plantation. Nevertheless, their sameimportance carried over to the post-liver transplantation period as a portalto inject contrast medium noninva-sively to evaluate for surgical compli-cations in the postoperative period.

Postoperative liver transplant com-plications that can be detected with T-tube cholangiography in the posttrans-plantation period include anastomoticand nonanastomotic complications.Anastomotic complications includeanastomotic leaks and anastomotic ste-noses. Early postoperative anastomoticcomplications are probably technical innature (8–16). Nonanastomotic earlypostoperative complications of the bileducts are peripheral bile duct injurysecondary to a primary, usually vas-

e cholangiography in an adult liver transpthe end-to-end biliary anastomosis only. Nally through the sphincter of Oddi and inter transplant recipient after the intravenor the intrahepatic ducts. The image showslled intrahepatic ducts.

cular, complication (15,16). In other

words, the biliary complications are ac-tually cholangiographic findings thatare complementary to other findings(liver enzymes, liver biopsy, clinical pic-ture, quality of hepatic graft) that helpsupport the diagnosis of a vascular in-jury such as early postoperative hepaticartery thrombosis, increased cold isch-emia time or reperfusion injury of thegraft. As a result, we deem an adequatepostoperative cholangiogram as onethat evaluates for a truly technical sur-gical complication (anastomotic compli-cation) and an adequate representativeevaluation of the intrahepatic peripheralbile ducts to evaluate for a diffuse bili-ary process that may reflect a globalgraft complication. Our definition of anadequate representative evaluation ofthe intrahepatic ducts is visualization ofthe entire right-sided (anterior and pos-terior segments) peripheral bile ducts.This is because it is not uncommon tofind nonfilling of the left-sided periph-eral bile ducts (�20% in this study).

Surgically placed small-bore biliarytubes come in various shapes and

t recipient. Contrast medium opacifies thee of the graft biliary anatomy is opacified.e small bowel (B). (b) Fluoroscopic image

administration of 4 mg of morphine. Thee moderate narrowing at the anastomosis

ub lanes onic o th

liv usve th

sizes. Some are actually T shaped and


others end blindly without a particularconfiguration or anchoring device/shape. In addition, there are severalways and places to place them. Theycan be passed through bowel or liveror can just lay in the infrahepatic peri-toneum (1,2). Some are placed in a Uconfiguration where the tube has anentering and an exiting limb. In addi-tion to the varying placements andconfigurations of the cholangiogramtubes, there are varying liver trans-plants and varying biliary anastomo-ses (1–3). The tube and anastomoticvariations are so varied that they can-not all be adequately described in thelimited scope of this article. However,radiologists should be aware of the sur-gical anatomy that they are dealing withand whether morphine can hypotheti-cally work in particular tube placementsand particular surgical anatomy. Oneobvious example is the futility in aug-menting T-tube cholangiography withmorphine in patients with biliary entericanastomoses because there is no sphinc-ter of Oddi to contract (recipient sphinc-ter is excised).

The key point in this study is thatintravenously administered morphinecan render 92% of nondiagnostic T-tube cholangiograms (via patent, well-positioned tubes) diagnostic examina-tions. Because this was a retrospectivestudy and the administration of mor-phine was operator-dependent, ninecases were not given morphine. If we,hypothetically, add these cases, con-sidering the same morphine successrate (92%), we would add an addi-tional eight (8/9) successful cases.Therefore, the potential to improve thediagnostic yield of T-tube cholangio-grams via well-positioned, patent T-tubes is 14% (increase of 14%) from adiagnostic yield of 85% to a diagnosticyield of 99%.

There are substantial limitations tothis retrospective study. Due to its ret-rospective nature, we can only hy-pothesize the full extent of how suc-cessful morphine can increase theoverall diagnostic yield of T-tube chol-angiograms. As a result, we can onlyextrapolate that morphine can increasethe diagnostic yield to up to 99% (seeabove). Conversely, we do not usuallypractice the use of table tilt and patientrotation/repositioning to help redis-tribute contrast medium within thebiliary tract by attempting to allow
contrast medium to fill the more de-
pendent parts of the biliary tract. Thedegree of effectiveness of this practicein adequately visualizing the biliarytract is unknown. Furthermore, not allangiography suites have tilt tables toplace patients in the Trendelenburg po-sition. If these maneuvers are effective,they would probably reduce the need togive intravenous morphine---particu-larly when considering, although minor,the additional cost and medicationsrisks of administering morphine.

Also due to the retrospective natureof the study, we do not know howmany T-tubes were blocked and re-opened by using forced saline injec-tion and 0.014–0.018-inch wires. Thisis a not-uncommon practice at our in-stitution. We do not know to whatdegree individual operators went to inorder to clear blocked tubes. Becauseof this, the frequency of blocked T-tubes in this study should be takenwith this perspective. In the same re-spect, malpositioned T-tubes can bemanipulated (usually pulled back ifthey have migrated inward) to obtaina diagnostic examination. We do notknow how many tubes were manipu-lated and to what extent individualoperators went to obtain a diagnosticexamination. However, neither thefrequency, nor the degree of success atresolving tube blockages and malposi-tions, affect the core results of thisstudy—which is the effectiveness ofintravenous morphine in augmentingT-tube cholangiograms given the ap-propriate surgical anatomy.

The current retrospective study didnot show any effect of patient age orsex on T-tube cholangiogram ade-quacy (P � .05) (Table 1). Similarly,the time of the T-tube cholangiogramand the presence of anastomotic steno-ses appeared to not have any effect onthe diagnostic adequacy of postopera-tive T-tube cholangiograms (P � .05)(Table 1). In addition, when compar-ing patients who received intravenousmorphine and those who did not inthe 22 patients with poor intrahepaticbile duct filling, there is no statisticallysignificant difference (P � .05) in theage, sex, and time from transplanta-tion between the two groups (Table 2).

However, there is a statistically sig-nificant difference (P � .012) betweenthe degree of anastomotic stenoses inrecipients with inadequate filling whosubsequently received intravenous
morphine and those who did not (Ta-
ble 2). However, despite the fact thatthis is a retrospective study, we firmlyassert that there is no bias as to whoreceived morphine. The divide be-tween who received and who did notreceive morphine was operator prefer-ence. Furthermore, the use of mor-phine, with its ability to increase con-trast backfilling of the biliary tract,may have helped diagnose anasto-motic stenoses. In other words, thecontinued lack of filling (without mor-phine administration) may have had theoperator(s) overlook anastomotic steno-ses at T-tube cholangiography. This canbe seen in the seven cases that receivedintravenous morphine, where the con-trast medium had reached but did notdelineate the anastomosis before mor-phine administration. One of the sevencases had a significant stenosis that wasdelineated due to morphine augmenta-tion; without morphine, 10% of signifi-cant stensoses in this series (cohort)would not have been diagnosed.

The frequency of cholangitis afterT-tube cholangiography in this studyis 2.8% (4/145) despite prophylacticantibiotic administration. This is con-sistent with the rate of post--T-tubecholangitis in studies on nontrans-plant patients who had a cholangitisrate of 1.2%–5.8% and a sepsis rate of0%–1.2% without antibiotic prophy-laxis (7,17,18). These studies recom-mended not to give prophylactic anti-biotics (7,17,18). However, due to thesusceptibility of liver transplant recip-ients for infections, particularly in theearly postoperative period, we giveprophylactic antibiotics. To the best ofour knowledge, we do not have refer-ences for T-tube cholangiography intransplant recipients for a direct com-parison. However, we recommend anti-biotic prophylaxis for T-tube cholan-giography in liver transplant recipients.The lack of cholangitis in the 13 patientswho received morphine is most likelycoincidental; despite the small samplesize, we do not believe that intravenousmorphine would significantly increasethe frequency of post--T-tube cholan-giography cholangitis.

In conclusion, intravenous mor-phine was successful in 92% of cases inopacifying the biliary tract withoutcomplications in liver transplant recip-ients with choledocho-choledochosto-mies and intact sphincter of Oddi. Inthis study, intravenous morphine in
cases of well-positioned T-tubes in-


creased the diagnostic yield of T-tubecholangiograms from 85% to 93% andhas the potential, if applied to allcases, to increase the diagnostic yieldto 99%. Furthermore, sex, age, timefrom transplantation, and presence ofsignificant anastomotic stenoses donot affect the opacification of the bili-ary tract via T-tubes.

References1. Blumgart LH, D’Angelica M, Jarnagin

WR. Biliary-enteric anastomoses. In:Blumgart LH, ed. Surgery of the Liver,Biliary Tract, and Pancreas. 4th ed.Philadelphia: Saunders, 2006; 455–474.

2. Saad WEA, Orloff MC, Davies MG,Waldman DL, Bozorgzadeh A. Post-transplant biliary and vascular surgicalanatomy. Tech Vasc Interv Radiol 2007;10:172–190.

3. Saad WEA. Unconventional cholan-giography. Tech Vasc Interv Radiol2008; 11:90–101.

4. Krishnamurthy S, Krishnamurthy GT.Cholecystokinin and morphine pharma-cological intervention during 99mTc-HIDA cholescintigraphy: a rational ap-proach. Semin Nucl Med 1996; 26:16–24.

5. Pedersen SA, Oster-Jorgensen E, Kra-glund K. The effects of morphine on
biliary dynamics: a scintigraphic study
with 99mTc-HIDA. Scand J Gastroen-terol 1987; 22:982–986.

6. van der Linden W, Kempi V. Fillingof the gallbladder as studied by com-puter-assisted Tc-99m HIDA scintigra-phy: concise communication. J NuclMed 1984; 25:292–298.

7. Dellinger EP, Kirshenbaum G, Wein-stein M, Steer M. Determinants of ad-verse reaction following postoperativeT-tube cholangiogram. Ann Surg 1980;191:397–403.

8. Zajko AB, Campbell WL, Bron LM,Schade RR, Koneru B, Van Thiel DH.Diagnostic and interventional radiol-ogy in liver transplantation. Gastroen-ter Clin North Am 1988; 17:105–143.

9. McDonald V, Matalon TAS, Patel SK,et al. Biliary strictures in hepatic trans-plantation. J Vasc Interv Radiol 1991; 2:533–538.

10. Zajko AB, Sheng R, Zetti GM, Madar-iaga JR, Bron KM. Transhepatic bal-loon dilation of biliary strictures inliver transplant patients: a 10-year ex-perience. J Vasc Interv Radiol 1995; 6:79–83.

11. Terblanche J, Allison HF, NorthoverJMA. An ischemic basis for biliarystrictures. Surgery 1983; 94:52–57.

12. Northover JMA, Terblanche J. A newlook at the arterial supply of the bileduct in man and its surgical implica-
tions. Br J Surg 1979; 66:379–384.
13. Saad WEA, Saad NEA, Davies MG,et al. transhepatic balloon dilation ofanastomotic biliary strictures in livertransplant recipients: the significanceof a patent hepatic artery. J Vasc IntervRadiol 2005; 16:1221–1228.

14. Saad WEA, Davies MG, Saad NEA, et al.Transhepatic dilation of anastomotic bil-iary strictures in liver transplant recipi-ents utilizing a combined cutting andconventional balloon protocol: technicalsafety and efficacy. J Vasc Interv Radiol2006; 17:837–843.

15. Orons PD, Sheng R, Zajko AB. Hepaticartery stenosis in liver transplant recipi-ents: prevalence and cholangiographicappearance of associated biliary compli-cations. AJR Am J Roentgenol 1995; 165:1145–1149.

16. Saad WEA. Trans-hepatic managementof post-operative anastomotic strictures.Tech Vasc Interv Radiol 2008; 11:143–153.

17. Sheen-Chen SM, Cheng YF, Chou FF,Lee TY. Postoperative T-tube cholan-giography: is routine antibiotic pro-phylaxis necessary? A prospective,controlled study. Arch Surg 1995; 130:20–23.

18. Pitt HA, Postier RG, Cameron JL.Postoperative T-tube cholangiography:is antibiotic coverage necessary? Ann
Surg 1980; 191:30–34.

intravenous morphine for augmentation of postoperative t-tube cholangiograms in liver transplant...

Documents