IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

The value of postoperative measurement of amylase inabdominal drainage fluid after pancreatic surgeryF Safi, S Bakathir, M Taha, T Lange, H El Salhat, F Branicki

ABSTRACTIntroduction: Clinical symptoms accompaniedby a continuous increase of amylaseconcentration in abdominal drainage fluid andchange in color of drainage fluid may indicatethe presence of fistula or leakage. Aims: Toinvestigate the clinical relevance and utility ofpost­operative (PO) monitoring of amylase andlipase estimations in the serum and abdominaldrainage fluid following pancreatic surgery.Methods: Seventy patients (37 males, 33females) who underwent duodenum­preservingpancreatic head resection [n = 12 (GI)], pyloruspreserving Whipple’s procedure [n = 39 (GII)],segmental resection of the body of the pancreas[n = 4 (GIII)] and pancreas tail/body resection[n = 15] were enrolled in the study prospectively.In G I, II and III (n = 55) duct mucosaanastomosis with the remnant of distal pancreaswas fashioned. The serum amylase and lipaselevels and levels of amylase in drainage fluidwere measured pre­operatively and from PO day1, until removal of the drain. Only 32 patientsreceived subcutaneous octreotide, 100 µg threetimes daily for five days. Results: Elevation of

serum amylase (≥100 IU/l) was found in 20/52(38%) patients following pancreatic surgery.The elevated amylase levels returned to normalwithin four days. Abdominal drainage fluidamylase values were found increased in 19/47(40%) of patients. All elevated levels returned tonormal by the tenth post­operative day. Thecolor of abdominal drainage fluid was sero­sanguinous in all cases. No clinical pancreaticfistula or anastomotic leakage was evident (0/55patients), all patients were discharged home.Conclusion: Documentation of transientelevation of serum amylase and abdominaldrainage fluid amylase levels did not appear tobe of clinical significance.Keywords: Pancreatic anastomosis, Leakage,Fluid amylase level.

*********Safi F, Bakathir S, Taha M, Lange T, Salhat HE, BranickiF. The value of postoperative measurement of amylasein abdominal drainage fluid after pancreatic surgery.International Journal of Hepathobiliary and PancreaticDiseases 2012;2:31–37.Article ID: 100009IJHPDFS2012

*********doi:10.5348/ijhpd­2012­9­OA­7

INTRODUCTIONThe three fundamental techniques for reconstructionto enable pancreatic exocrine secretion after proximalpancreas resection are end to side pancreatico­jejunostomy, end to side pancreatico­gastrostomy andend to end invaginating anastomosis.

ORIGINAL ARTICLE OPEN ACCESS

1F Safi, 2S Bakathir, 3M Taha, 4T Lange, 2H El Salhat, 1FBranickiAffi l iations: 1Department of Surgery, United Arab EmiratesUniversity; 2Department of Surgery, Tawam Hospital, AlAin, UAE; 3Department of Internal Medicine (DivisionGastroenterology), Tawam Hospital, Al Ain, UAE; 4WeimerHospital, Germany.Corresponding Author: Dr. Farouk Safi, ProfessorDepartment of Surgery, Faculty of Medicine & HealthSciences, United Arab Emirates University, P.O. Box1 7666, Al Ain, United Arab Emirates; Email :f.safi@uaeu.ac.ae

Received: 27 December 2011Accepted: 03 Apri l 201 2Published: 08 November 201 2

Safi et al. 31

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

Safi et al. 32

The mortality rate after pancreatic resection hasdecreased to less than 5% but reported morbidityremains high [1]. Even today, following resection,morbidity ranges from 6–57%. Fistula rate afterresection is 0–20% with a morbidity of 0–13% [1]. Themain cause of serious morbidity and mortality ispostoperative anastomotic leakage due to intra­peritoneal release of enterokinases and activation ofpancreatic enzymes that lead to sepsis and hemorrhagiccomplications. Failure of the remnant proximalpancreatic duct closure after left sided pancreaticresection causes milder complications as compared toanastomotic failure of an entero­entero anastomosis butit can result in fistula formation or an intra­abdominalfluid collection [1].The failure in the integrity of an anastomosisbetween the remnant pancreas and the gastrointestinaltract can result in a fistula with leakage of pancreaticsecretions outside the abdominal cavity. This may occurwithout significantly affecting the general condition ofthe patient or it may prolong in­hospital treatment orlead to discharge of the patient with the fistula untilspontaneous closure occurs. Failure of anastomoticintegrity may well give rise to intra­abdominal fluidcollections, sepsis, hemorrhage, peritonitis and death.Strasberg et al. classified pancreatic anastomotic failureinto five grades depending on the clinical coursefollowing this complication. Pancreatic anastomoticfailure grade I requires normal post­operative therapywithout any need for additional treatment; in grade IIthere is need for pharmacological support; grade IIIrequires interventional therapy; grade IV requiressurgical and intensive care treatment and grade V isdeath [2].The International Study Group for Pancreatic FistulaDefinition reviewed several previous scores/definitionsof pancreas anastomotic failure on the basis of twoparameters; 1) daily output in ml, 2) and amylaseconcentration in the drained abdominal fluid, and 3) theduration of the fistula. Suspicion of and diagnosis offistula depends on monitoring the output of theabdominal drainage system, the concentration ofamylase in drainage fluid more than three times theserum concentration after third post­operative day,color of the fluid, general condition of the patient, andradiologic documentation. The Group also proposed anA, B or C clinical grading system: a) well patient, b)often well, and c) appearing ill [1].Theoretically, the drainage fluid should be rich inamylase when pancreatic fluid leaks from the pancreaticstump and anastomotic failure. Therefore, ‘amylase rich’drainage has been one of the most popular definitions ofpancreatic duct leakage used in the literature [3]. Thepurpose of are study was: 1) to investigate the value ofmeasurement of amylase concentration in serum andabdominal drainage fluid in diagnosing anastomotic orstump failure after pancreatic surgery and to determineif a correlation exists between the dynamics of post­operative changes in amylase concentration in drainagefluid with clinical outcome, 2) identify whether highconcentrations of amylase in the serum or in abdominal

drainage fluid postoperatively or during deterioratingclinical status of the patient are essential for thediagnosis of anastomotic failure, 3) to follow up thepatients with high abdominal drainage/serum amylaselevels in order to document whether the patientdeveloped surgical complications in the form ofanastomotic or stump failure and, 4) to evaluate thevalue of octerotide in reducing failures.

MATERIALS AND METHODSA prospective study was conducted which includedall patients undergoing pylorus preserving proximalpancreaticoduodenectomy (PD), duodenum preservingpancreas head resection (DPPHR), segmental resectionof the pancreas with anastomosis (SR) and distalpancreatectomy (DP) by the same surgeon betweenJanuary 2000 ­ July 2011. Data collected includedpatient demographics, type of operation carried out,hospital mortality, morbidity, and the need forreoperation and readmission.Instructions were given to nurses and residentsregarding the need to measure the serum andabdominal drainage fluid amylase and lipaseconcentrations daily starting from the first postoperativeday and to inject octereotide subcutaneously in a dosageof 3x100 µg; 3x200 µg or 3x300 µg for variable periods.All patients who underwent pancreatic resection andanastomosis had two drains (without suction), oneplaced in a sub­hepatic location, the second adjacent tothe pancreatic anastomosis. A single abdominal drainwas used for patients undergoing DP and was placedadjacent to the pancreatic stump. The decision toremove the drains was dependent on the levels ofamylase recorded postoperatively, the color of theabdominal fluid and the general clinical condition of thepatient. The upper limit of the reference range ofamylase in our laboratory was 100 IU/L.Surgical technique: Pancreaticojejunostomosiswas performed by the same surgeon (FS) in all patientsas follows. A retro­colic jejunal limb was broughtthrough a window in the transverse colon, an end to sidepancreatico­jejunal anastomosis was fashioned in twolayers; a duct to mucosa anastomosis was made usingPDS® 5/0 or 6/0 interrupted sutures; all knots wereplaced outside the lumen. The second layer involved apancreas parenchyma­seromuscular interrupted suturesusing 4/0 PDS®. This technique was performedregardless of the diameter of the pancreatic duct.Following distal pancreatectomy, the pancreatic ductwas closed separately using prolene® 4/0 U shapedsutures, followed by parenchyma­parenchyma suturingusing interrupted prolene® 4/0 sutures.

RESULTSResults are summarized in table 1–4. The tablesshow the number of patients with serum amylase valuesof more than 100 IU/L. There was missing data for

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

Safi et al. 33

amylase drainage fluid later in the post­operative coursewhen drains had been removed or no samples were sentfor testing to the laboratory. The study included 70patients. There male to female ratio was 1.12:3. Meansage of all patients was 57.4 years.There were no post­operative deaths. Morbidity notrelated to pancreatic surgery was observed in 18patients (26%). Four patients (5.7%) required re­operation for control of hemorrhage. None of thepatients developed any clinical signs of fistula and oranastomotic/stump leakage. No abdominal collectionwas documented based on clinical signs and symptoms.Pancreatic surgery caused post­operativeamylasaemia in about 40% patients without any clinicalsigns of pancreatitis. The elevated serum amylase valuesreturn to normal, less than 100 IU/L, within one weekof operation. High serum values were as high as 10 foldthe normal levels (Table 2). No patient exhibited clinicalsigns of acute pancreatitis.More than 50% operated patients showed elevatedamylase values in the abdominal drainage fluidcollections. The concentration of the amylase was ashigh as 200 fold the normal value. All increased levelsreturned to normal within 7–10 days. The color ofcollected body fluid was initially serosanginous thenserous. The observed amylase levels did not play anyrole in the decision as to when the drain can beremoved. The decision to remove the drain was made

Table 1: Patient demographic data (n = 70).

considering the color of the fluid, normalizing ordecreasing amylase concentrations and the generalcondition of the patient. Neither abdominal sepsis norintra­abdominal collections of fluid were documented inany patients (Tables 3, 4).

DISCUSSIONThe mortality rate following duodeno­pancreatectomy has been reported between 3%–5% insome large series [3, 4]. A mortality rate of zero hasbeen reported in some series of more than a hundredpatients [5–7]. The major cause of mortality is sepsisand/or hemorrhage resulting from failure ofpancreatico­jejuno or gastrostomosis [8]. To preventthis life threatening complication after surgery, variousmodifications for pancreatico­enteric reconstructionhave been proposed like pancreatico­jejunostomy orpancreatico­gastrostomy, invagination or duct tomucosa anastomosis, stented or nonstentedanastomosis, end to end or end to side anastomosis andthe use of fibrin glue. However, no universal consensushas been reached as to which particular variation ofpancreatico­enteric reconstruction is safer and lessprone to anastomotic failure. Randomized controlledtrials and metaanalysis showed no difference in leakrates between pancreatico­gastro or jejunostomosis [9,

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

Safi et al. 34

Table 2: Serum amylase levels pre­ and post­pancreatic resection (n = 70).

Table 3: Postoperative abdominal drainage amylase levels (subhepatic drain).

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

Safi et al. 35

Table 4: Postoperative abdominal drainage amylase Levels (drain adjacent to pancreatic anastomosis).

10]. Observational clinical studies demonstrate that thesurgeon’s experience and preferences play to influenceoperative outcome. Having become familiar withprinciples relating to pancreatic surgery the sametechnique has been used to connect the remnantpancreas with the jejunum and employs an end to sideduct mucosa double layer with interrupted 4/0 PDS®sutures. No evident leakage in our study with oursurgical technique provides evidence for the selection ofthis safe anastomotic procedure [11].Pancreatic fistula remains a significant cause of post­operative morbidity and mortality in patients undergoingpancreatic head resection [12]. Traditionally, thediagnosis of a fistula has been made on the basis ofamylase rich drainage fluid or radiological evidence ofdisruption of the pancreatic anastomosis. The reportedincidence of post­operative pancreatic fistula is between0–30% [1]. The true incidence is, however, unknownbecause of the varying definition as to what constitutes apost­operative pancreatic fistula [1]. Anastomotic orstump failure after pancreatic surgery is a serious surgicalcomplication. It occurs in immediate or easypostoperative period causing deviation from the normalpost­operative course requiring careful management [2].Whether the diagnosis of this complication can be madeon clinical grounds or by measurement of amylase levelsmerits consideration.We analyzed data relating to amylase levels in serumand abdominal drainage fluid to determine how usefulthese are to identify anastomotic or stump failure.Hyperamylasaemia has been reported after manysurgical procedures. Occasionally, the origin is salivaryamylase and this does not reflect injury of the pancreas.

In such cases amylase levels return to normal within oneweek post­operatively [13]. Measurement of serumamylase after pancreatic surgery is not mandatory andshould be done only if there is clinical suspicion ofpancreatitis. In all our patients with hyperamylasaemiano clinical features of pancreatitis were documented.The volume of drainage fluid per day has beenselected as one indicator of pancreatic fistula. In thereview reported by Shinchi et al. [14], some authorschose 50 ml/day as a ‘cut­off’ value, others selected 30ml/day and in some publications the volume is notdefined, as in our analysis. We decided to remove drainsafter three to four days according to the color of thedrainage fluid and the general condition of the patient.A drain producing serous fluid with a normal amylaseconcentration can be removed independent of thevolume of drainage fluid.High concentrations of amylase in body fluids do notalways indicate anastomotic or stump failure. Themeasurement of amylase concentration only oncepostoperatively in the third, fifth, seventh or tenth post­operative day in body fluid is also of no value for thediagnosis and definition of anastomotic fistula oranastomotic failure. Some authors have chosen two fold,two and half fold, three fold, five fold or absolute valuesof ≥1000, 2000, or 5000 for amylase concentration ascut off values for the recognition of this complication[15, 16]. Our study is consistent with the reports fromHiroyuki et al. [14] and Yi­Ming Shyr et al. [17] thatthere are patients without any clinical signs of leakage orsymptoms who have high amylase value in body fluidseven seven to 10 days post­operatively. Perhaps, in somepatients, pancreatic juice leaks from needle hole sites or

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

Safi et al. 36

from the anastomosis itself, which has not fully healed.It may be that amylase levels are high and do actuallyindicate leakage of pancreatic fluid but this issubclinical in nature without overt manifestation ofleakage like large intra­abdominal collections orfeatures of systemic sepsis.The dynamic and course of amylase concentration ineach patient during the complete post­operative periodis of more diagnostic value than the estimation ofamylase on a particular post­operative day. In all ourpatients without leakage the high amylaseconcentration returned to normal levels. Yi­Ming Shyret al. [17], reported that all patients without leakageshowed a continuous decrease in amylase levels. Inaddition, patients without leakage reported by Hiroyukiet al had a reduction in amylase concentration and nearnormalization by the 13th post­operative day.Decreasing concentrations of amylase re­increasingwere found only in patients with clinical leakage [14]with very high values beyond the 10th post­operativeday.Neither the isolated estimation of amylase in serum,nor the volume or amylase concentration of abdominaldrainage fluid, on a single post­operative day appears tobe indicative of the complications of fistula,anastomotic failure or leakage. More important is thedynamic of daily evaluation of the amylaseconcentration, which after 10 days, is more likely to beof diagnostic value. Suspicion of these complicationsmay be raised clinically on daily patient rounds. Thediagnosis can be confirmed, better delineated andfollowed by additional laboratory tests, imaging studiesand monitoring of the course of measuring amylaseconcentrations in body fluids [18, 19].

CONCLUSIONIncrease in serum amylase concentration andabdominal drainage fluid without clinical symptomsafter pancreatic surgery on a single post­operative daydid not appear to be indicative of the complications offistula, anastomotic failure or leakage. Only thecontinuous increase beyond tenth post­operative daysin drain fluid amylase concentration with change incolour of drainage fluid and appearance of clinicalsymptoms indicate the above mentioned complications.

*********AcknowledgementsThe authors would like to acknowledge the contributionfrom Mr. Abdulla Jamal for his technical help with thepreparation of the manuscript.Author ContributionsF Safi – Major Contribution in conception and design,Acquisition of data, Analysis and interpretation of data,Substantial contribution in drafting the article, criticalrevision of the article and submission, Correspondingauthor and final approval of the version to be published

S Bakathir – Minor contribution in Conception, andAcquisition of data, Minor contribution in drafting thearticle, Minor contribution in final approval of theversion to be publishedM Taha – Minor contribution in Conception, andAcquisition of data, Minor contribution in drafting thearticle, Minor contribution in final approval of theversion to be publishedT Lange – Major Contribution in conception and design,Acquisition of data, Analysis and interpretation of data,Minor contribution in drafting the article, Minorcontribution in final approval of the version to bepublishedH El Salhat – Minor contribution in Conception, andAcquisition of data, Minor contribution in drafting thearticle, Minor contribution in final approval of theversion to be publishedF. Branicki – Minor contribution in Conception, andAcquisition of data, Minor contribution in drafting thearticle, Major contribution in final approval of theversion to be publishedGuarantorThe corresponding author is the guarantor ofsubmission.Conflict of InterestAuthors declare no conflict of interest.Copyright© F Safi et al. 2012; This article is distributed under theterms of Creative Commons Attribution 3.0 Licensewhich permits unrestricted use, distribution andreproduction in any means provided the originalauthors and original publisher are properly credited.(Please see www.ijhpd.com/copyright­policy.php formore information.)

REFERENCES1. Bassi C, Dervenis C, Butturini G, et al. Postoperativepancreatic fistula: an international study group(ISGPF) definition. Surgery 2005;138(1):8–13.2. Strasberg SM, Linehan DC, Clavien PA, Barkun JS.Proposal for definition and severity grading ofpancreatic anastomosis failure and pancreaticocclusion failure. Surgery 2007;141(4):420–6.3. Yeo CJ, Cameron JL, Sohn TA, et al. Six hundredfifty consecutive pancreaticoduodenectomies in the1990s: pathology, complications, and outcomes. AnnSurg 1997;226(3):248–57.4. Sutton CD, Garcea G, White SA, et al. Isolated Roux­loop pancreaticojejunostomy: a series of 61 patientswith zero postoperative pancreaticoenteric leaks. JGastrointest Surg 2004;8(6):701–5.5. Trede M, Schwall G, Saeger HD. Survival afterpancreatoduodenectomy. 118 consecutive resectionswithout an operative mortality. Ann Surg1990;211(4):447–58.6. Aranha GV, Hodul PJ, Creech S, Jacobs W. Zeromortality after 152 consecutivepancreaticoduodenectomies with

IJHPD 201 2;2:31 –37.www.ijhpd.com

IJHPD – International Journal of Hepatobil iary and Pancreatic Diseases, Vol. 2, 201 2. ISSN – [2230 - 901 2]

pancreaticogastrostomy. J Am Coll Surg2003;197(2):223–1.7. Cameron JL, Pitt HA, Yeo CJ, Lillemoe KD, KaufmanHS, Coleman J. One hundred and forty­fiveconsecutive pancreaticoduodenectomies withoutmortality. Ann Surg 1993;217(5):430–5.8. Duffas JP, Suc B, Msika S, et al. A controlledrandomized multicenter trial ofpancreatogastrostomy or pancreato­jejunostomyafter pancreatoduodenectomy. Am J Surg 2005;189(6):720–9.9. Bassi C, Falconi M, Molinari E, et al. Reconstructionby pancreaticojejunostomy versus pancreatico­gastrostomy following pancreatectomy: results of acomparative study. Ann Surg 2005;242(6):767–1.10. Wente MN, Shrikhande SV, Müller MW, et al.Pancreaticojejunostomy versuspancreaticogastrostomy: systematic review andmeta­analysis. Am J Surg 2007;193(2):171–83.11. You D, Jung K, Lee H, Heo J, Choi S, Choi D.Comparison of different pancreatic anastomosistechniques using the definitions of the InternationalStudy Group of Pancreatic Surgery: a singlesurgeon's experience. Pancreas2009;38(8):896–902.12. Büchler MW, Friess H, Wagner M, Kulli C, WagenerV, Z'Graggen K. Pancreatic fistula after pancreatichead resection. Br J Surg 2000;87(7):883–9.13. Murakami Y, Uemura K, Hayasidani Y, et al. A softpancreatic remnant is associated with increaseddrain fluid pancreatic amylase and serum CRP levelsfollowing pancreatoduodenectomy. J GastrointestSurg 2008;12(1):51–6.14. Shinchi H, Wada K, Traverso LW. The usefulness ofdrain data to identify a clinically relevant pancreaticanastomotic leak after pancreaticoduodenectomy? JGastrointest Surg 2006;10(4):490–8.15. Okano K, Kakinoki K, Suto H, et al. Persisting ratio oftotal amylase output in drain fluid can predictpostoperative clinical pancreatic fistula. JHepatobiliary Pancreat Sci 2011 May 19.16. Molinari E, Bassi C, Salvia R, et al. Amylase value indrains after pancreatic resection as predictive factorof postoperative pancreatic fistula: results of aprospective study in 137 patients. Ann Surg2007;246(2):281–7.17. Shyr YM, Su CH, Wu CW, Lui WY. Does drainagefluid amylase reflect pancreatic leakage afterpancreaticoduodenectomy? World J Surg2003;27(5):606–10.18. Strasberg SM, Drebin JA, Mokadam NA, et al.Prospective trial of a blood supply­based techniqueof pancreaticojejunostomy: effect on anastomoticfailure in the Whipple procedure. J Am Coll Surg2002;194(6):746–58.19. Grobmyer SR, Rivadeneira DE, Goodman CA,Mackrell P, Lieberman MD, Daly JM. Pancreaticanastomotic failure after pancreaticoduodenectomy.Am J Surg 2000;180(2):117–20.

Safi et al. 37