Objective Splenectomy with or without distal pancreatectomy may be necessary at time of cytoreductive surgery to achieve complete cytoreduction in advanced ovarian cancer. However, these procedures have been associated with peri-operative morbidity. The aims of this study were to determine the incidence of distal pancreatectomy among patients undergoing splenectomy during cytoreductive surgery for advanced ovarian cancer and to determine the incidence, management, treatment, and prognosis of patients with post-operative pancreatic fistula.
Methods Retrospective cohort study of all consecutive patients with FIGO stage IIIC-IVB ovarian, fallopian tube, or primary peritoneal cancer who underwent splenectomy with or without distal pancreatectomy, during primary, interval, or secondary cytoreductive surgery between January 2007 and December 2017. All histologic subtypes were included; patients with borderline ovarian tumor and those undergoing emergency surgery were excluded from analysis. Univariate analyses for survival were generated by Kaplan–Meier survival curves and log-rank (Mantel–Cox) tests for statistical significance. Patients who underwent surgery for recurrence were excluded from survival analysis. Inter-group statistics were performed using Student’s t-test for continuous variables, and chi-square test and Fisher’s exact test for categorical variables.
Results A total of 156/804 (19.4%) women underwent splenectomy, and of these 22 (14.1%) patients had distal pancreatectomy. Of patients who underwent splenectomy only, 2/134 (1.5%) developed grade B post-operative pancreatic fistula and 6/22 (27.3%) patients who underwent distal pancreatectomy developed grade B and C post-operative pancreatic fistula. Five (83.3%) of six of these patients were symptomatic. Distal pancreatectomy patients had a higher risk of developing post-operative pancreatic fistula when compared with patients who underwent splenectomy only (63.7% vs 9.7%, p=0.0001). Median length of hospital stay was longer in patients with post-operative pancreatic fistula: 16.5 (range 7–38) days compared with 10 (range 7–15) days (p=0.019). There was no progression-free survival (p=0.42) and disease-specific survival (p=0.33) difference between patients undergoing splenectomy with or without distal pancreatectomy.
Conclusion Clinically relevant post-operative pancreatic fistula is a relatively frequent complication (27.3%) following distal pancreatectomy and it is a possible complication after splenectomy only (1.5%).
- ovarian cancer
- surgical oncology
- digestive system fistula
Statistics from Altmetric.com
A total of 19.4% of stage IIIC-IV ovarian cancers undergoing cytoreductive surgery require splenectomy.
Clinically relevant post-operative pancreatic fistula following distal pancreatectomy is a frequent complication (27.3%).
A small proportion of patients who undergo splenectomy only develop clinically relevant post-operative pancreatic fistula (1.5%).
Ovarian cancer represents the most lethal gynecologic cancer, with 295 414 new cases estimated and 184 799 deaths in 2018 worldwide.1 Its disproportionately high death rate can be attributed to the advanced stage at presentation with more than 70% of women being diagnosed with International Federation of Obstetrics and Gynecology (FIGO) stage III or IV.2 The prognostic importance of complete cytoreduction either in the primary surgery setting or following neoadjuvant chemotherapy at interval surgery is universally accepted,3 4 with decreasing evidence on the role of cytoreductive surgery for the treatment of platinum-sensitive recurrent ovarian cancer, particularly in view of the recently published results from a randomized controlled trial.5
Metastatic disease in the upper abdomen in ovarian cancer has historically been one of the main barriers to achieving a complete cytoreduction,6 7 but improvement in surgical technique alongside peri-operative care have enabled this barrier to be overcome, balancing ultra-radical surgery with acceptable morbidity and mortality.8 As result, there has been a change in surgical paradigm with an increasing number of upper abdominal procedures with time.9 Among these procedures, distal pancreatectomy is required in about 20%–41% of patients undergoing splenectomy to achieve complete cytoreduction.10 11 However, distal pancreatectomy can be associated with high risk of post-operative complications, particularly post-operative pancreatic fistula. Management of post-operative pancreatic fistula can prolong length of hospital stay and may be a major determinant of morbidity associated with a significant risk of delay in the start of adjuvant chemotherapy. Very few studies have specifically addressed the incidence and management of post-operative pancreatic fistula after distal pancreatectomy in advanced ovarian cancer10–12 and the impact on progression-free survival or overall survival is not clear. Additionally, the validated definition and the grading of post-operative pancreatic fistula has recently been modified by the International Study Group of Pancreatic Surgery (ISGPS)13 and this new classification has yet to be applied to the ovarian cancer population.
The primary objective of this study was to determine the incidence of distal pancreatectomy among patients undergoing splenectomy in a large series of surgically treated advanced ovarian cancer patients from a United Kingdom tertiary cancer center and to determine the incidence, management, treatment, and prognosis of patients with post-operative pancreatic fistula. The secondary objective was to perform a survival analysis of patients who underwent splenectomy with or without distal pancreatectomy.
In this retrospective cohort study all consecutive patients with FIGO stage IIIC-IVB ovarian, fallopian tube, or primary peritoneal cancer or recurrent disease who had undergone splenectomy with or without distal pancreatectomy, during primary cytoreductive surgery, interval cytoreductive surgery, or secondary cytoreductive surgery at the Northern Gynecological Oncology Centre, Gateshead, United Kingdom (from January 2007 to December 2017) were identified from the departmental electronic database, after institutional review board approval and the institution’s audit department registration (Number 1315/2017). All histologic subtypes were included; however, patients with borderline ovarian tumor and those undergoing emergency surgery were excluded from analysis.
Clinicopathologic records were accessed in conjunction with morbidity data from the prospectively collected departmental database.
All patients underwent pre-operative computed tomography (CT) scan of chest, abdomen, and pelvis, measurement of serum CA125, and discussion at the gynecologic oncology multidisciplinary team. Decision to primary versus interval cytoreductive surgery was made by gynecologic oncologists at the multidisciplinary team meeting after review of the CT scan by one radiologist dedicated to gynecologic oncology (>10 years’ experience) and after assessment of the patients’ performance status. Cytoreductive surgery was undertaken via laparotomy and completeness of cytoreduction was defined by the greatest dimension (in cm) of the largest residual lesion and classified according to the completeness of cytoreduction in three groups: no gross macroscopic residual disease (R0), 0–1 cm, or >1 cm (these last two groups were considered as R>0 for survival analysis).
Distal pancreatectomy was performed when bulky metastatic tumor was found on the distal pancreas or in the splenic hilum inseparable from the pancreatic tail. The technique to resect and seal the distal pancreas was variable according to individual surgeon’s preference and included: (a) linear stapler with or without reinforcing suture to oversew the staple line, (b) LigaSure Impact bipolar device (Medtronic, Minneapolis, Minnesota, USA), and (c) hand-sewn suture only. The use of hemostatic agents including Surgicel/Evicel (Ethicon, Somerville, NJ, USA), or Floseal (Baxter, Deerfield, IL, USA) was also according to individual preference.
The team performing surgery was composed of one senior gynecologic oncology consultant with expertise in advanced ovarian cancer surgery. He trained the other consultants and trainees in the department. During the time period of the study, another experienced consultant joined the department and was part of the team for a limited period. As part of departmental protocol, patients undergoing splenectomy with or without distal pancreatectomy had a measurement of drain fluid amylase level from left upper quadrant drain, when this was inserted, on post-operative days 3–5.
Extent of surgery was classified by the surgical complexity score14 and intra-operative complications were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) system (version 4.03).15 The post-operative complications were prospectively collected and defined as any adverse event related to operative treatment occurring within 30 days of surgery, including death. These were graded according to the Clavien–Dindo grading system.16
Post-operative pancreatic fistula was defined according to the 2016 update of International Study Group of Pancreatic Surgery (ISGPS)13 as any measurable volume of drain fluid on or after post-operative day 3 with amylase level >3 times the upper limit of normal serum amylase activity, associated with a clinically relevant development/condition related directly to the post-operative pancreatic fistula. This is graded in three grades according to the clinical outcome (Figure 1): ‘biochemical leak’, defined as leak of pancreas-derived enzyme-rich fluid with no clinical impact (therefore no longer considered a true post-operative pancreatic fistula); ‘grade B’, defined as a change in post-operative management (drains left in place >3 weeks or repositioned through endoscopic or percutaneous procedures); and ‘grade C’, defined as a post-operative pancreatic fistula which requires reoperation or leads to single- or multi-organ failure and/or mortality attributable to the post-operative pancreatic fistula. Following splenectomy, all patients were given vaccines and lifelong penicillin-based antibiotics.
The time to adjuvant chemotherapy was defined as the number of days from post-operative day 1 to the first day of adjuvant chemotherapy. Progression-free survival was calculated as the time in months from the date of the cytoreductive surgery to the date of first progression, last follow-up, or death. Disease-specific survival was defined as the time in months from the date of the cytoreductive surgery to the date of the last follow-up or death related to the ovarian, fallopian tube, or primary peritoneal cancer; patients who died of a different cause were censored from the survival analysis. Univariate analyses for disease-specific survival and progression-free survival were generated by Kaplan–Meier survival curves17 and log-rank (Mantel–Cox) tests for statistical significance. Follow-up data were collected until March 2020. Patients who underwent surgery for recurrence were excluded from survival analysis.
Inter-group statistics were performed using Student’s t-test for continuous variables, and chi-square test and Fisher’s exact test for categorical variables. All statistical tests were performed using SPSS statistical software program, version 26.0 (SPSS Inc., Chicago, IL, USA). All statistical tests were two-sided, and differences were considered significant at a level of p<0.05.
Overall Patient Characteristics
Over the 11-year study period, 804 patients underwent cytoreductive surgery for FIGO stage IIIC-IVB and recurrent ovarian, fallopian tube, or primary peritoneal cancer. Of these, 156 (19.4%) underwent splenectomy, with or without distal pancreatectomy. In 28/156 (17.9%) patients the pre-operative CT scan specifically reported disease related to the spleen or the distal pancreas. Clinicopathologic characteristics of the patients and the tumors are shown in Table 1.
A total of 89/156 (57.0%) patients underwent primary cytoreductive surgery, 50/156 (32.1%) patients interval cytoreductive surgery (after 3–4 cycles of neoadjuvant chemotherapy), and 17/156 (10.9%) patients secondary cytoreductive surgery. The residual disease at the end of the cytoreductive surgery was <1 cm in 144 patients, 92.3% of the overall cohort. The remaining 12 (7.7%) patients had residual disease >1 cm due to: intra-operative patient anesthetic instability in 3 (25.0%) patients, 15 mm maximum diameter disease on small bowel serosa in 6 (50.0%) patients, retro-pancreatic disease in 1 (8.3%) patient, and disease in porta hepatis in 2 (16.7%) patients.
There was a clear trend with time of increasing performance of splenectomy and distal pancreatectomy along with increasing rates of no macroscopic residual disease, that is, number of splenectomies increased from 11 in 2007/2008 to 45 in 2015/2016. Overall, 3/156 (1.9%) patients died in the post-operative period, of whom two died within 30 days and the third on post-operative day 38.
Splenectomy Only Patients
Of 134 patients who underwent splenectomy only, 13 (9.7%) patients developed post-operative pancreatic fistula: 11 (8.2%) had biochemical leak, 2 (1.5%) had grade B post-operative pancreatic fistula, and no patient had grade C post-operative pancreatic fistula. Fifty-five (41.0%) patients did not have measurement of drain fluid amylase on post-operative days 3–5 either because no drain was inserted at the end of the surgery or because the drain fluid amylase was not measured. A total of 16 patients (11.9%) had grade 3 intra-operative complications. There was no intra-operative grade 4/5 complications. A total of eight women experienced grade 3 (6.0%) and three patients experienced grade 4 (2.2%) post-operative complications. One patient (0.7%) died within 30 days of surgery of a pulmonary embolism on post-operative day 1.
Splenectomy with Distal Pancreatectomy Patients
A total of 22 patients (14.1% of splenectomy patients and 2.7% of entire stage IIIC or stage IV cohort undergoing cytoreductive surgery) had distal pancreatectomy performed at the time of cytoreductive surgery. There were no cases of distal pancreatectomy performed without splenectomy. All distal pancreatectomies were performed by gynecologic oncology surgeons with support from an upper gastrointestinal surgeon in five cases (22.7%). The median age in the distal pancreatectomy group was 69 (range 32–77) years.
Of patients undergoing distal pancreatectomy, 13 (59.1%) patients had a surgical complexity score in the range 4–7 and 9 (40.9%) patients had a score ≥8. No patient experienced intra-operative complications of grade ≥3. Only four (18.2%) patients who underwent distal pancreatectomy had a post-operative complication ≥3 including one 30 days post-operative death. These were: one patient with a right pleural effusion requiring drainage, one patient with a pancreatic fistula requiring drain re-insertion, one patient had an episode of asystole due to arrhythmia in intensive care unit, and one patient died of sepsis from a bowel anastomosis leak (unrelated to the splenectomy/distal pancreatectomy).
The technique to seal the pancreatic stump during the distal pancreatectomy was hand-sewn suture in 6 (27.3%) patients, mechanical stapler in 3 (13.6%) patients, LigaSure Impact bipolar device (Medtronic, Minneapolis, MN, USA) in 6 (27.3%) patients, and stapler with reinforcing suture to oversew the staple line in 1 (4.5%) patient. In six patients the surgeon did not document the technique used in the operation notes. Hemostatic agents were used in 11 patients to the pancreatic stump (7 cases Surgicel/Evicel and 4 cases Floseal). Pancreatic tissue was confirmed histologically in all 22 patients.
Patients who had distal pancreas removed had a significantly higher risk of developing post-operative pancreatic fistula when compared with patients who underwent splenectomy only (63.7% vs 9.7%, p=0.0001). The characteristics of the 22 patients who underwent intentional removal of distal pancreas and the differences between women who developed or did not develop post-operative pancreatic fistula are demonstrated in Table 2. In one patient the left upper quadrant drain was not positioned at the end of the surgery and in another patient the drain fluid amylase was not measured on post-operative days 3–5. Fourteen of twenty-two (63.7%) patients with distal pancreatectomy developed post-operative pancreatic fistula, of which 4 (18.2%) patients had grade B and 2 (9.1%) patients had grade C.
Of the four patients who underwent distal pancreatectomy and developed grade B post-operative pancreatic fistula two had post-operative pancreatic fistula-related sepsis and two had persistent drainage of fluid with abnormally elevated amylase for more than 3 weeks (with no clinical consequence). One of the patients with grade C post-operative pancreatic fistula had persistent leakage of pancreatic fluid for more than 3 weeks with multi-organ failure related to a left upper quadrant collection resulting in bleeding from the pancreas stump/splenic vessels on post-operative day 43 requiring angiographic embolization. The second patient with grade C post-operative pancreatic fistula developed post-operative pancreatic fistula-related sepsis leading to multi-organ failure and subsequently to death 38 days after the surgery; this condition had developed on a background of chest infection, poor mobility, and poor nutritional status.
The median drain fluid amylase measured on post-operative days 3–5 in patients who developed post-operative pancreatic fistula was 1949.5 (range 244–51 229) U/L. The median value of serum amylase in these women was 29 (range 10–144) U/L. Acute pancreatitis was not reported in any case. When a drain was re-inserted in the left upper quadrant due to grade B or C post-operative pancreatic fistula, the median time for drain re-insertion was 27 (range 10–32) days from the surgery. There was no association between any of the techniques used to seal the pancreatic stump after distal pancreatectomy and the development of grade B or C post-operative pancreatic fistula (p=1.0). Range of time of left upper quadrant drain placement was significantly longer in patients who developed post-operative pancreatic fistula (post-operative pancreatic fistula, median 12 (range 4–31) days; no post-operative pancreatic fistula, median 5 (range 3–6) days; p=0.008).
Median length of hospital stay was longer in patients with post-operative pancreatic fistula: 16.5 (range 7–38) days compared with 10 (range 7–15) days (p=0.019). Median time from surgery to the first cycle of chemotherapy was 60 (range 27–79) days in patients with post-operative pancreatic fistula in comparison to 46 (range 35–56) days in patients without (p=0.185). Patients with grade B/C post-operative pancreatic fistula were symptomatic in 5/6 (83.3%) cases. In three of these patients the initial presentation was temperature >38.0°C, and in two patients left upper quadrant abdominal pain. Four patients with post-operative pancreatic fistula were treated with octreotide (somatostatin analog) and one patient was administered octreotide and total parenteral nutrition.
Median follow-up for the entire cohort was 23 (range 2–92) months. The median progression-free survival was 13 months (95% CI 10.7 to 15.2) for the splenectomy only patients and 14 months (95% CI 7.4 to 20.5) for the distal pancreatectomy patients (Figure 2A). Median disease-specific survival was 33 months (95% CI 25.1 to 40.8) for the splenectomy only group, while the median was not reached in the distal pancreatectomy group (Figure 2B). The 5-year estimated disease-specific survival rate was 23.1% in splenectomy only patients and 53.7% in distal pancreatectomy patients. No difference was seen when progression-free survival and disease-specific survival of patients undergoing splenectomy only were compared with patients undergoing splenectomy with distal pancreatectomy (progression-free survival, p=0.42; disease-specific survival, p=0.33).
The survival analysis performed according to the completeness of cytoreduction in the 139 (89.1%) patients undergoing primary cytoreductive surgery or interval cytoreductive surgery demonstrated a significant difference in both progression-free survival and disease-specific survival when no residual disease was compared with any residual disease (progression-free survival, p=0.017; disease-specific survival, p=0.005). Median progression-free survival was 17 months (95% CI 12.0 to 21.9) and 11 months (95% CI 7.7 to 14.2), respectively, for patients who had residual disease 0 and >0 at the end of the surgery (Figure 3A). Median disease-specific survival was greater than 39 months (95% CI 33.5 to 44.5) and 23 months (95% CI 17.9 to 28.0), respectively, for patients who had residual disease 0 and >0 at the end of the surgery (Figure 3B).
A substantial number of patients with advanced ovarian cancer present with upper abdominal disease. Of these, about 14%–20% require splenectomy to accomplish complete cytoreduction to no gross residual disease.18 Distal pancreatectomy is performed even more rarely as part of cytoreductive surgery in ovarian cancer, and to date very few series describing outcomes related to this procedure have been published.10–12 19
Post-operative pancreatic fistula is an important complication that requires proactive assessment during the post-operative period. Based on the new ISGPS classification,13 post-operative pancreatic fistula is diagnosed with the measurement of drain fluid amylase on or after post-operative day 3. Therefore, the importance of a left upper quadrant drain insertion in the splenic/pancreatic bed at the time of the surgery becomes evident. In our series of distal pancreatectomy, all but one patient had a left upper quadrant drain inserted at the time of the surgery and in all but one of the remaining cases the drain fluid amylase was measured on post-operative days 3–5. In these patients, post-operative pancreatic fistula occurred in 63.7%, the majority being Grade A defined as biochemical leak.
According to the 2016 ISGPS update,13 biochemical leak is no longer considered a ‘true’ fistula, being a finding with no clinical relevance. In our distal pancreatectomy series, we reported four grade B post-operative pancreatic fistulae and two grade C post-operative pancreatic fistulae, giving a 27.3% incidence of ‘true’ post-operative pancreatic fistula. This incidence is in keeping with previous reports where the occurrence of clinically relevant post-operative pancreatic fistulae ranged from 22.2% to 29%.10–12 It is also important to note that patients who underwent splenectomy only developed post-operative pancreatic fistula in 9.7% of cases, but only 1.5% had a ‘true’ post-operative pancreatic fistula (grade B). This signifies that some pancreatic tissue had either been removed or been disrupted at the time of spleen removal. The morbidity associated with splenectomy in our series may be considered acceptable (9.0% with grade 3–4 post-operative complications and 1.9% peri-operative mortality). The overall morbidity in terms of intra-operative and post-operative adverse events is comparable with published series.11 18
Symptoms of relevant post-operative pancreatic fistula were represented by left upper quadrant pain and temperature >38.0°C. Subsequent measurements of drain fluid amylase in patients with elevated levels of drain fluid amylase on post-operative days 3–5 are necessary to correlate these symptoms to the pancreatic leak and take action accordingly. Furthermore, the level of white cell counts on post-operative day 3 did not differ in patients who did or who did not experience post-operative pancreatic fistula. Most women in our cohort had leukocytosis on post-operative day 3; leukocytosis is often seen after splenectomy and previous reports show that this, as well as platelet-to-white blood cell ratio, may not be a marker of infection when present as an isolated finding.20 21 Patients with post-operative pancreatic fistula had a significantly longer period of hospitalization; this was due to the time for completion of treatment (antibiotics, octreotide, total parenteral nutrition) and the time for normalization/reduction of drain fluid amylase. Conversely, there was no significant delay in the time to start chemotherapy, which supports the feasibility of distal pancreatectomy in order to achieve complete cytoreduction.
Treatment of post-operative pancreatic fistula includes a number of strategies. A drain to the left upper quadrant should be re-inserted if previously removed in order to prevent abscess formation. CT scan monitoring is critical in addition to use of broad-spectrum antibiotics as prophylaxis to prevent development of post-operative pancreatic fistula-related sepsis. A controversial issue is diet modification and the administration of total parenteral nutrition. A randomized trial showed that enteral nutrition is superior to total parenteral nutrition in terms of closure rate and time to closure of post-operative pancreatic fistula.22 Also, the use of somatostatin analogs remains contentious.23 24
Survival comparison between our cases of splenectomy with distal pancreatectomy and splenectomy only showed no difference in both progression-free survival and disease-specific survival, further supporting the role of radical surgery to achieve complete cytoreduction in advanced ovarian cancer patients. Moreover, in agreement with literature studies, the progression-free survival and disease-specific survival of patients who received complete cytoreduction to no gross residual disease was significantly better than for patients with any residual disease.25 26
The main strength of our study is that, to our knowledge, this is the largest series from a single referral center of cases of splenectomy with or without distal pancreatectomy in advanced ovarian cancer. Furthermore, it is the first gynecologic oncology study to adopt the latest classification of ISGPS,13 and it provides novel data on the proportion of cases undergoing splenectomy also requiring distal pancreatectomy. A limitation of this study was that the pathologists did not comment in the pathology report whether there was presence of pancreatic tissue in the specimens submitted as ‘splenectomy’ only. Another limitation is represented by the fact that 41.0% of splenectomy patients did not have measurement of drain fluid amylase on post-operative days 3–5, so we could have underestimated the diagnosis of biochemical leak; nevertheless, post-operative information on all these patients was recorded, and no clinically relevant post-operative pancreatic fistula occurred.
In conclusion, 19.4% of stage IIIC and stage IV ovarian cancers undergoing cytoreductive surgery require splenectomy in order to achieve satisfactory cytoreduction rates. Of these splenectomy cases, distal pancreatectomy is required in 14.1% of patients. Clinically relevant post-operative pancreatic fistula following distal pancreatectomy is a relatively frequent complication (27.3%). A small proportion of patients who undergo splenectomy only also develop clinically relevant post-operative pancreatic fistula (1.5%). Therefore, even in cases of splenectomy without distal pancreatectomy, a left upper quadrant drain should be inserted at the time of surgery to diagnose and treat post-operative pancreatic fistula promptly. Overall, distal pancreatectomy is a feasible procedure with acceptable post-operative morbidity which allows the achievement of cytoreduction in advanced ovarian cancer patients.
The Authors would like to thank the surgical team at Northern Gynecological Oncology Centre.
Presented at Preliminary results of this study were presented at the 20th Biennial International Meeting of the European Society of Gynaecological Oncology (ESGO 2017), Vienna, Austria, November 4-7, 2017.
Contributors NB: conceptualization, data curation, methodology, writing original draft. PK: data curation, methodology, conceptualization. VG: data curation, methodology, review and editing, RLOD: data curation, methodology, review and editing, SR: data curation, methodology, review and editing. RN: conceptualization, methodology, review and editing, writing original draft.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.