Splenic vein thrombosis and pancreatic fistula after minimally invasive distal pancreatectomy

Original article

Splenic vein thrombosis and pancreatic fistula after minimallyinvasive distal pancreatectomy

C. M. Kang1,3, Y. E. Chung2,3, M. J. Jung1,3, H. K. Hwang1,3, S. H. Choi1,3 and W. J. Lee1,3

Departments of 1Surgery and 2Radiology, Yonsei University College of Medicine, and 3Pancreaticobiliary Cancer Clinic, Institute of Gastroenterology,Severance Hospital, Seoul, KoreaCorrespondence to: Dr W. J. Lee, Department of Surgery, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Korea(e-mail: [email protected])

Background: This study aimed to investigate the clinical relevance of splenic vein thrombosis (SVT) inthe splenic vein remnant following minimally invasive distal pancreatosplenectomy (DPS).Methods: Medical records of patients who underwent laparoscopic or robotic distal pancreatectomy(DP) with or without splenectomy between January 2006 and August 2012 were reviewed. Rates of SVTand clinically relevant postoperative pancreatic fistula (POPF) were compared in a group of patientsundergoing DPS and a group having spleen-preserving DP.Results: Seventy-nine patients had minimally invasive DP, of whom 38 (48 per cent) developed SVT inthe splenic vein remnant. DPS was associated with POPF (P = 0·001) and SVT (P < 0·001). SVT lengthwas closely related to the amount of peripancreatic fluid collection (P = 0·025) and POPF (P = 0·045). Ina comparison of splenic vessel-sacrificing, spleen-preserving DP and DPS, postoperative platelet countwas significantly higher in the DPS group (P < 0·001). In addition, grade of SVT (P = 0·092) and POPF(P = 0·065) tended to be associated with DPS, suggesting that SVT may be related to both splenectomyand POPF.Conclusion: Minimally invasive DPS is associated with SVT and POPF. Preservation of the spleenshould be considered when treating patients with benign and borderline malignant tumours of the distalpancreas.

Paper accepted 30 September 2013Published online 10 December 2013 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.9366

Introduction

Laparoscopic distal pancreatectomy (DP) is currentlythe standard surgical approach for benign or borderlinemalignant tumours of the pancreas, and is consideredsafe and effective1–3. Spleen-preserving procedures arepreferred over routine combined splenectomy in viewof the role of the spleen in immune defence4,5.However, splenic vessel-conserving, spleen-preserving DPis difficult and time-consuming, and it is not alwayspossible to preserve the spleen. Warshaw6 introducedthe alternative technique of spleen-preserving DP withsegmental resection of both the splenic artery andvein. This is much easier than the splenic vessel-conserving approach, but carries the risk of splenicinfarction7 and perigastric variceal bleeding8. However,more than 20 years of experience have suggested that

this splenic vessel-sacrificing, spleen-preserving DP issafe9. With recent advances in surgical techniqueand devices, including the da Vinci Surgical System(Intuitive Surgical, Sunnyvale, California, USA), spleen-preserving procedures can be performed safely andeffectively10.

A recent case of a patient who underwent electivelaparoscopic distal pancreatosplenectomy (DPS) anddeveloped a clinically relevant postoperative pancreaticfistula (POPF)11, associated with splenic vein thrombosis(SVT) in the remnant distal part of the splenic vein, andpartial portal vein thrombosis on postoperative follow-upcomputed tomography (CT), formed the impetus for thepresent study. Specifically, this study aimed to investigatewhether there is a clinical association between minimallyinvasive DP and the development of postoperative SVTand clinically relevant POPF.

2013 BJS Society Ltd BJS 2014; 101: 114–119Published by John Wiley & Sons Ltd

Splenic vein thrombosis and minimally invasive distal pancreatectomy 115

Methods

The medical records of patients who underwent minimallyinvasive (laparoscopic or robotic) DP between January2006 and August 2012 were reviewed. The institutionalreview board approved this retrospective study protocol.Patients with malignant lesions of the pancreas and thosewho had extended left-sided pancreatectomy requiringdivision of the neck of the pancreas and vascular controlat the origin of the splenic vessels12 were excluded. Inminimally invasive DP, the reasons for preserving or notpreserving the spleen may differ according to surgeon pref-erence, surgical technique and departmental philosophy.In general, the present authors try to preserve the spleenwhen performing minimally invasive DP in patients withbenign and borderline malignant tumours, and a splenicvessel-conserving, spleen-preserving DP is usually per-formed in these patients. In patients with excessive bloodloss and longer operating times, the procedure is convertedto a splenic vessel-sacrificing, spleen-preserving DP. Rel-atively large pancreatic tumours abutting splenic vessels,associated with chronic pancreatitis, are a common indica-tion for splenic vessel-sacrificing, spleen-preserving DP13.Splenectomy is usually indicated when: too much bleedingoccurs; splenic perfusion is impaired after splenic vessel-sacrificing, spleen-preserving DP; or the patient requiresa reduced duration of surgery because of co-morbidity.

All patients routinely underwent abdominal CT duringpreoperative evaluation and before discharge on day7 after surgery. Images were obtained with a 16- or64-channel multidetector CT scanner (SOMATOMSensation 16 or Sensation 64; Siemens Medical Solutions,Forchhein, Germany). Using a bolus tracking technique,pancreatic parenchymal (or late arterial) phase scanningwas performed with a scan delay of 23 s once the abdominalaorta reached 100 Hounsfield units. The portal venousphase was obtained with a scan delay of 25 s after the endof the previous phase. CT parameters were: 0·5-s rotationtime, 120 kV, 240 mA, 0·6-mm beam collimation, beampitch 1, and 3-mm slice thickness.

The classification system for determination of splenicvessel patency has been described previously14,15. Briefly,the patency of the splenic artery and vein were classifiedinto three grades according to the degree of stenosis:intact (none), partial occlusion or thrombosis (partial),and total occlusion or not identified (complete). Tworadiologists reached consensus in assessing the degree ofSVT on retrospective review of CT images. Postoperativeperipancreatic fluid collections were estimated using thelongest diameter of the fluid collection measured on axialimages and the total number of 3-mm thickness CT imagesections covering the whole extent of the fluid collection.

Complete blood count, serum amylase and peritonealfluid amylase levels were checked on days 1, 3, 5and 7 after surgery. Postoperative morbidities weredefined as follows: POPF11 (clinically relevant, grade Bor above, amylase concentration in the drainage fluidmore than 3 times the upper limit of normal serumamylase after the third postoperative day, requiringadditional medical and interventional treatment); intra-abdominal abscess (follow-up abdominal CT showinglocalized intra-abdominal fluid collection, accompanyingair bubbles, or peripheral wall enhancement with feverand leucocytosis (white blood cell count above 10 000/µl));postpancreatectomy haemorrhage16 (grade B or above,requiring blood transfusion, interventional radiology, orsurgical intervention associated with bloody dischargefrom the drain or haematoma around the surgical fielddefined on follow-up CT); wound infection (presenceof pus requiring wound opening); intestinal obstruction(characteristic imaging finding of mechanical obstructionassociated with clinical presentation such as vomiting,abdominal distension, absence of flatus, or abdominaldiscomfort requiring nasogastric tube decompression).

Statistical analysis

Continuous variables are expressed as mean(s.d.). Statisticaldifferences were evaluated by χ2 analysis and Student’s ttest, performed as a paired analysis when necessary. Linearregression analysis was also used for statistical assessment ofassociations between continuous variables, such as lengthof splenic vein thrombosis and size of peripancreaticfluid collections. All statistical analyses were performedusing SPSS for Windows version 20.0 (IBM, Armonk,New York, USA). P < 0·050 was considered statisticallysignificant.

Results

During the study interval, 79 patients of mean age50(15) years underwent minimally invasive DP for benignand borderline malignant tumours of the pancreas.Twenty-six patients (33 per cent) were women and 53(67 per cent) were men. The most common pathologicalindication was a solid pseudopapillary tumour (18, 23 percent) (Table 1). Mean tumour diameter was 3·0(1·7) cm,and length of resected pancreas 7·8(2·8) cm. A spleen-preserving DP was performed in 45 patients: splenicvessel-conserving in 36 patients (46 per cent), and splenicvessel-sacrificing in nine (11 per cent). A total of 59 patients(75 per cent) had conventional laparoscopic DP, and 20(25 per cent) had robot-assisted DP. The robotic surgical

2013 BJS Society Ltd www.bjs.co.uk BJS 2014; 101: 114–119Published by John Wiley & Sons Ltd

116 C. M. Kang, Y. E. Chung, M. J. Jung, H. K. Hwang, S. H. Choi and W. J. Lee

Table 1 Pathological diagnoses

No. of patients (n = 79)

Solid pseudopapillary tumour 18 (23)Mucinous cystic neoplasm 16 (20)Serous cystic tumour 15 (19)Intraductal papillary mucinous neoplasm 14 (18)Neuroendocrine tumour 11 (14)Chronic pancreatitis 3 (4)Intrapancreatic accessory spleen 1 (1)Epithelial cyst 1 (1)

Values in parentheses are percentages.

system was used more frequently in spleen-preserving DPcompared with the laparoscopic approach (laparoscopic:28 for spleen-preserving DP versus 31 for DPS; robotic:17 versus 3 respectively; P = 0·002). Mean postoperativehospital stay was 11(7) days. Clinically relevant POPF wasnoted in 22 patients (28 per cent); in eight patients (10 percent) this was associated with an intra-abdominal abscess.There were no cases of postpancreatectomy haemorrhageand no perioperative hospital mortality. Prophylacticlow molecular weight heparin (LMWH) was used in 16patients (20 per cent).

Effect of spleen preservation on pancreatic fistulaand splenic vein thrombosis

All splenic veins were shown to be patent on preoperativeCT. Following minimally invasive DP, 38 (48 per cent) of79 patients were found to have SVT in the distal splenicvein remnant. The incidence of SVT and postoperativethrombocytosis was significantly increased in the DPSgroup (P < 0·001). Two of the patients undergoing DPShad signs of additional thrombosis in the intrahepaticportal system. In addition, the DPS group exhibited morefrequent POPF (P = 0·001) compared with the spleen-preserving DP group (Table 2).

Clinical implications of splenic vein thrombosisand its association with pancreatic fistula

There was a close relationship between SVT length andthe maximum size of the peripancreatic fluid collection(R2 = 0·076, P = 0·025). The rate (P = 0·022) and severity(P = 0·027) of SVT were significantly associated withPOPF. There was no significant difference in the maximumsize of peripancreatic fluid collections between the studygroups, although patients with POPF had an increasedSVT length (P = 0·045) (Table 3).

Table 2 Effect of spleen preservation on postoperative pancreaticfistula, splenic vein thrombosis and platelet count

Spleen-preservingDP (n = 45)

DPS(n = 34) P†

POPF 0·001No 39 18Yes 6 16

SVT* < 0·001No 33 6Yes 11 27

Mean(s.d.) platelet count(× 103/µl)

< 0·001‡

Preop. 267·6(103·0) 243·8(61·3)Postop. 296·5(148·7)§ 567·1(246·4)¶

*Analysis excluded two patients with no immediate postoperativefollow-up computed tomography results. DP, distal pancreatectomy;DPS, distal pancreatosplenectomy; POPF, postoperative pancreaticfistula (defined according to Bassi et al.11); SVT, splenic vein thrombosis.†χ2 test, except ‡paired Student’s t test. §P = 0·080; ¶P < 0·001 versuspreop. value (paired Student’s t test).

Table 3 Association between splenic vein thrombosis andpostoperative pancreatic fistula

POPF

No (n = 58) Yes (n = 21) P‡

SVT† 0·022§No 33 6Yes 23 15

SVT grade† 0·027¶None 33 7Partial 14 6Complete 9 8

SVT length (mm)* 6·7(10·2) 15·2(17·5) 0·045Maximum PFC size (mm)* 52·2(90·5) 51·3(21·7) 0·965LOS (days)* 7·9(3·0) 18·1(9·5) < 0·001

*Values are mean(s.d.). †Analysis excluded two patients with noimmediate postoperative follow-up computed tomography results. POPF,postoperative pancreatic fistula; SVT, splenic vein thrombosis; PFC,peripancreatic fluid collection; LOS, length of hospital stay. ‡Student’s ttest, except §χ2 test and ¶χ2 test with linear-by-linear association.

Effect of splenic vessel-sacrificing, spleen-preserving policy

Splenic vessel-sacrificing, spleen-preserving DP and DPSwere compared to determine the impact of splenectomyon SVT and POPF. The number of patients with splenicvessel-sacrificing, spleen-preserving DP was limited (9patients), and the incidence of SVT was not significantlydifferent between the two groups (P = 0·181). However,the SVT tended to be larger (P = 0·056) and moresevere (P = 0·092) in the DPS group. In addition, moreextensive peripancreatic fluid collections (P = 0·039)and significantly higher postoperative platelet counts



(P < 0·001) were noted following DPS. Finally, POPFtended to be associated more with DPS than with splenicvessel-sacrificing, spleen-preserving DP (P = 0·065) (datanot shown).

Discussion

In DP, the spleen can be preserved both safely andeffectively using a robotic surgical system10, and follow-upCT shows that the patency of conserved splenic vessels islargely preserved in these procedures15. Laparoscopic DPShas been associated with higher-grade complications, pro-longed hospital stay and diminished quality of life5. Thus,every effort should be made to preserve the spleen in DPfor benign and borderline malignant pancreatic tumours.The present study further corroborates the potentialadverse impact of splenectomy during DP on postoperativeoutcomes such as SVT in the splenic vein remnant. Theassociation with clinically relevant POPF suggests anotherpotential rationale for spleen preservation in DP.

Few studies to date have looked at postoperativeSVT and its clinical importance following minimallyinvasive DPS; most17–19 have considered SVT following

laparoscopic ‘splenectomy’. It should be noted that therate of asymptomatic SVT following splenectomy hasbeen reported as 5–52 per cent20,21. Impaired clearanceof prothrombotic factors, reduced splenic venous flowand postoperative thrombocytosis have all been proposedas potential mechanisms for postoperative SVT aftersplenectomy20,21. In the present study, the incidence ofSVT following minimally invasive DP was high (48 percent). In particular, 27 (79 per cent) of 34 patients whohad minimally invasive DPS exhibited signs of SVT inthe splenic vein remnant. Moreover, SVT size, severity ofSVT and peripancreatic fluid collections were much moreextensive in the DPS group than in the spleen-preservingDP group. Comparison of the subgroup of patients whounderwent splenic vessel-sacrificing, spleen-preservingDP with those undergoing DPS suggests that SVT mayhave resulted from a concomitant splenectomy, althoughthe number of patients was small.

There was no association between the prophylacticuse of LMWH and the rate of SVT (data not shown).This was also demonstrated in a recent randomizedclinical trial17 investigating the impact of anticoagulationon the incidence of splenic or portal vein thrombosis

DPS

Reduced blood flow in splenic vein

Stasis and turbulence in distal part of splenic vein SVT

Congestive ischaemia Wound healing failure

PFCClinically relevant

POPF

Long-term drain placement Splenectomy-associated impaired

immunological function Infection

Splenectomy-induced thrombocytosis

Fig. 1 Proposed mechanism of splenic vein thrombosis (SVT) and clinically relevant postoperative pancreatic fistula (POPF) in distalpancreatosplenectomy (DPS). Combined splenectomy-induced thrombocytosis and haemodynamic changes may result in SVT in theremnant splenic vein after DPS, leading to a failure of the wound healing process in the pancreatic stump. An impaired immune systemas a result of splenectomy may add to the risk of infection. PFC, peripancreatic fluid collection


118 C. M. Kang, Y. E. Chung, M. J. Jung, H. K. Hwang, S. H. Choi and W. J. Lee

after laparoscopic splenectomy. However, the number ofpatients receiving prophylactic LMWH was too limited inthe present study to reach any conclusions.

The association between SVT and clinically relevantPOPF in minimally invasive DP is interesting. Recently,another study22 also reported a relatively high incidenceof POPF following DPS, although explanations have beenlacking. Based on the present observation, it is hypothe-sized that SVT may be causally related to clinically relevantPOPF in DPS. Splenectomy-induced haemodynamicchanges and combined thrombocytosis may thus result inSVT in the splenic vein remnant, which could influencethe wound healing process of the resected pancreas.Finally, subsequent peripancreatic fluid collections andimpaired host immunity can lead to clinically relevantPOPF (Fig. 1).

From the viewpoint of surgical strategies in routineDP for benign and borderline malignant tumours, thepresent study may suggest another rationale for spleenpreservation. In routine DP, it may be necessary totransect the pancreas at the neck (extended DP). Pancreatictransection at the neck is easier because of reducedparanchymal thickness, resulting in a lower risk of POPF23.Alternatively, the pancreatic division line may be decidedafter considering the anatomical relationship between thesplenic vein remnant and the inferior mesenteric vein,as the long splenic vein remnant that drains a residualpancreatic body with doubtful arterial vascularization maybe more prone to thrombosis.

An inherent limitation of the present study is itsretrospective observational design, based on institutionaldata collection over 15 years from a heterogeneous patientgroup. This is a critical drawback to the findings. Takingrecent technical improvements into consideration, thequality of surgical approaches and postoperative outcomesmay have been different in the early and later studyperiods. Spleen-preserving DP was performed morefrequently during the last 5 years of the study, whichundoubtedly contributed to bias. Based on the results ofthis retrospective cohort study, it is not clear whetherSVT is among the causative factors for POPF, or whetherPOPF itself increases subsequent SVT.

Disclosure

The authors declare no conflict of interest.

References

1 Xie K, Zhu YP, Xu XW, Chen K, Yan JF, Mou YP.Laparoscopic distal pancreatectomy is as safe and feasible as

open procedure: a meta-analysis. World J Gastroenterol 2012;18: 1959–1967.

2 Venkat R, Edil BH, Schulick RD, Lidor AO, Makary MA,Wolfgang CL. Laparoscopic distal pancreatectomy isassociated with significantly less overall morbidity comparedto the open technique: a systematic review andmeta-analysis. Ann Surg 2012; 255: 1048–1059.

3 Nigri GR, Rosman AS, Petrucciani N, Fancellu A, PisanoM, Zorcolo L et al. Metaanalysis of trials comparingminimally invasive and open distal pancreatectomies. SurgEndosc 2011; 25: 1642–1651.

4 Shoup M, Brennan MF, McWhite K, Leung DH, KlimstraD, Conlon KC. The value of splenic preservation with distalpancreatectomy. Arch Surg 2002; 137: 164–168.

5 Choi SH, Seo MA, Hwang HK, Kang CM, Lee WJ. Is itworthwhile to preserve adult spleen in laparoscopic distalpancreatectomy? Perioperative and patient-reportedoutcome analysis. Surg Endosc 2012; 26: 3149–3156.

6 Warshaw AL. Conservation of the spleen with distalpancreatectomy. Arch Surg 1988; 123: 550–553.

7 Baldwin KM, Katz SC, Espat NJ, Somasundar P.Laparoscopic spleen-preserving distal pancreatectomy inelderly subjects: splenic vessel sacrifice may be associatedwith a higher rate of splenic infarction. HPB (Oxford) 2011;13: 621–625.

8 Tien YW, Liu KL, Hu RH, Wang HP, Chang KJ, Lee PH.Risk of varices bleeding after spleen-preserving distalpancreatectomy with excision of splenic artery and vein. AnnSurg Oncol 2010; 17: 2193–2198.

9 Ferrone CR, Konstantinidis IT, Sahani DV, Wargo JA,Fernandez-del Castillo C, Warshaw AL. Twenty-three yearsof the Warshaw operation for distal pancreatectomy withpreservation of the spleen. Ann Surg 2011; 253:1136–1139.

10 Kang CM, Kim DH, Lee WJ, Chi HS. Conventionallaparoscopic and robot-assisted spleen-preservingpancreatectomy: does da Vinci have clinical advantages?Surg Endosc 2011; 25: 2004–2009.

11 Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C,Izbicki J et al.; International Study Group on PancreaticFistula Definition. Postoperative pancreatic fistula: aninternational study group (ISGPF) definition. Surgery 2005;138: 8–13.

12 Kang CM, Choi SH, Hwang HK, Kim DH, Yoon CI, LeeWJ. Laparoscopic distal pancreatectomy with division of thepancreatic neck for benign and borderline malignant tumorin the proximal body of the pancreas. J Laparoendosc AdvSurg Tech A 2010; 20: 581–586.

13 Choi SH, Kang CM, Kim JY, Hwang HK, Lee WJ.Laparoscopic extended (subtotal) distal pancreatectomy withresection of both splenic artery and vein. Surg Endosc 2013;27: 1412–1413.

14 Yoon YS, Lee KH, Han HS, Cho JY, Ahn KS. Patency ofsplenic vessels after laparoscopic spleen and splenicvessel-preserving distal pancreatectomy. Br J Surg 2009; 96:633–640.



15 Hwang HK, Chung YE, Kim KA, Kang CM, Lee WJ.Revisiting vascular patency after spleen-preservinglaparoscopic distal pancreatectomy with conservation ofsplenic vessels. Surg Endosc 2012; 26: 1765–1771.

16 Wente MN, Veit JA, Bassi C, Dervenis C, Fingerhut A,Gouma DJ et al. Postpancreatectomy hemorrhage (PPH): anInternational Study Group of Pancreatic Surgery (ISGPS)definition. Surgery 2007; 142: 20–25.

17 Wang H, Kopac D, Brisebois R, Sample C, Shapiro AM.Randomized controlled trial to investigate the impact ofanticoagulation on the incidence of splenic or portal veinthrombosis after laparoscopic splenectomy. Can J Surg2011; 54: 227–231.

18 Reoyo Pascual JF, Eldabe Mikhail A, Bayona Garcıa I, SecoGil JL. [Portal–splenic thrombosis after laparoscopicsplenectomy.] Cir Esp 2011; 89: 474–476.

19 Tran T, Demyttenaere SV, Polyhronopoulos G, Seguin C,Artho GP, Kaneva P et al. Recommended timing forsurveillance ultrasonography to diagnose portal splenic vein

thrombosis after laparoscopic splenectomy. Surg Endosc2010; 24: 1670–1678.

20 Stamou KM, Toutouzas KG, Kekis PB, Nakos S, Gafou A,Manouras A et al. Prospective study of the incidence and riskfactors of postsplenectomy thrombosis of the portal,mesenteric, and splenic veins. Arch Surg 2006; 141:663–669.

21 Ikeda M, Sekimoto M, Takiguchi S, Yasui M, Danno K,Fujie Y et al. Total splenic vein thrombosis afterlaparoscopic splenectomy: a possible candidate fortreatment. Am J Surg 2007; 193: 21–25.

22 Mekeel KL, Moss AA, Reddy KS, Mulligan DC, HaroldKL. Laparoscopic distal pancreatectomy: does splenicpreservation affect outcomes? Surg Laparosc Endosc PercutanTech 2011; 21: 362–365.

23 Pannegeon V, Pessaux P, Sauvanet A, Vullierme MP,Kianmanesh R, Belghiti J. Pancreatic fistula after distalpancreatectomy: predictive risk factors and value ofconservative treatment. Arch Surg 2006; 141: 1071–1076.


Documents

Splenic vein thrombosis and pancreatic fistula after minimally invasive distal pancreatectomy