Is prior cholecystectomy associated with decreased survival in patients with resectable pancreatic adenocarcinoma following pancreaticoduodenectomy?

The American Journal of Surgery (2011) 201, 519 –524

Clinical Science

Is prior cholecystectomy associated with decreased survival in patients with resectable pancreatic adenocarcinoma following pancreaticoduodenectomy? Suzanne C. Schiffman, M.D.a, Carrie K. Chu, M.D.b, Jaemin Park, B.S.b, Maria Russell, M.D.b, Steven Keilin, M.D.b, David A. Kooby, M.D.b, Charles R. Scoggins, M.D., M.B.Ch.B.a, Kelly M. McMasters, M.D., Ph.D.a, Robert C.G. Martin, M.D., Ph.D.a,* a

Division of Surgical Oncology, Department of Surgery, University of Louisville, and the James Graham Brown Cancer Center, Louisville, KY; bDivision of Surgical Oncology, Department of Surgery, Emory University School of Medicine, and the Winship Cancer Institute, Atlanta, GA, USA KEYWORDS: Pancreatic adenocarcinoma; Cholecystectomoy; Pancreaticoduedenectomy; Periampullary cancer

Abstract BACKGROUND: Patients with pancreatic cancer who present with biliary symptoms may undergo cholecystectomy and thus delay cancer diagnosis. We hypothesized that prior cholecystectomy leads to decreased overall survival in patients with pancreatic adenocarcinoma. METHODS: Retrospective study of hepatobiliary database. RESULTS: Three hundred sixty-five patients with a diagnosis of resectable periampullary pancreatic adenocarcinoma were identified. Eighty-seven patients underwent prior cholecystectomy. Median age (P ⫽ .48), body mass index (BMI) (P ⫽ .8), diabetes status (P ⫽ .06), American Society of Anesthesiologists (ASA) class (P ⫽ .22), stent placement (P ⫽ .13), operative time (P ⫽ .76), estimated blood loss (EBL) (P ⫽ .24), intraoperative transfusion (P ⫽ .91), portal vein resection (P ⫽ .25), LOS (P ⫽ .09) adjuvant therapy (P ⫽ .2), tumor size (P ⫽ .89), differentiation (P ⫽ .67), angiolymphatic invasion (P ⫽ .69), perineural invasion (P ⫽ 54), nodal metastasis (P ⫽ .43), complication rate (P ⫽ .75), and 30-day mortality (P ⫽ .58) were not statistically different between patients with previous cholecystectomy and those without. Median survival was 14 months for patients with a history of cholecystectomy and 16 months for those without (P ⫽ .25). Previous cholecystectomy was not a predictor of survival on Cox regression analysis. CONCLUSION: There was no difference in overall survival in patients with pancreatic cancer with prior cholecystectomy versus those without. © 2011 Elsevier Inc. All rights reserved.

* Corresponding author. Division of Surgical Oncology, Division of Surgical Oncology, University of Louisville, Norton, Healthcare Pavilion, 315 E. Broadway, Suite 303, Louisville, KY 40202. Tel.: ⫹1-5026293355; fax: ⫹1-502-629-3393. E-mail address: [email protected] Manuscript received December 18, 2009, accepted February 4, 2010

0002-9610/$ - see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjsurg.2010.02.004

Pancreatic cancer has a poor prognosis with an overall 5-year survival rate of less than 5%.1,2 This is largely attributed to nonspecific presenting symptoms and advanced disease at diagnosis. Lack of effective adjuvant therapy and systemic chemotherapy also contribute to the low survival rate. Recent studies failed to show any significant survival benefit with neoadjuvant therapy.3

520 Surgical resection offers the only chance of cure and is only possible if the patient is diagnosed at an early stage. Metastatic disease and local invasion into adjacent structures are contraindications to resection. Approximately 10% to 20% of patients present with pancreatic cancer amenable to resection.3,4 Multiple studies have demonstrated that pancreatic resection can be performed with an operative mortality of approximately 1% to 10%5–10 and the 5-year survival rate is approximately 5% to 15%.1,2,4,6,11 However, morbidity rates remain high at 30% to 60% and perioperative complications are common.2 Symptoms of pancreatic cancer may be falsely interpreted as those of biliary disease.12 Upper abdominal pain and jaundice can occur with diseases such as cholecystitis or cholelithiasis and thus result in delaying the diagnosis of pancreatic cancer. Laparoscopic cholecystectomy is the treatment of choice for the majority of patients with symptomatic cholelithiasis or cholecystitis. Some reports have indicated that a major drawback of laparoscopic surgery is the inability to palpate surrounding viscera.13–16 In theory, this could lead to the misdiagnosis or at least a delay in the diagnosis of abdominal malignancies. Patients who present with biliary symptoms may undergo cholecystectomy and subsequently delay cancer diagnosis. We sought to determine whether prior cholecystectomy leads to a decreased survival in patients with pancreatic adenocarcinoma.

The American Journal of Surgery, Vol 201, No 4, April 2011 margin (R0) was defined as microscopically tumor-free, whereas as microscopically positive margin was defined as R1. Clinical, radiologic, and histopathologic data were analyzed retrospectively. Postoperative complications were graded according to the 5-point grading scale and the highest severity level was recorded. Major complications were defined as grade 3 or higher. Thirty-day and 60-day mortality rates were calculated. Operative mortality was defined as patient death within 30 days of operation. Patients undergoing previous cholecystectomy were subdivided into recent versus remote cholecystectomy. Recent cholecystectomy was performed within the 2 years before pancreaticoduodenectomy. Patients who underwent cholecystectomy before resection were compared with patients who had not undergone cholecystectomy using Fischer exact, ␹2, and t test where appropriate. Statistical analysis was performed using JMP 4.0 (SAS Campus Drive, Cary, NC) and SPSS version 16 software (SPSS, Inc, Chicago, IL). Continuous variables were compared using Student t test and categorical variables were compared with ␹2 test. Survival was plotted using the method of Kaplan–Meier and compared using the log-rank test. P ⬍ .05 was considered a significant difference. Survival (in months) was measured from date of initial diagnosis until death or last follow-up visit. Cox regression was used to determine independent predictors of outcome. Multivariate analysis was performed with Cox proportional hazards model.

Methods The prospectively collected hepatobiliary databases at the University of Louisville and Emory University were queried and data combined for this retrospective study. Institutional review board approval was obtained from both institutions before analysis. Cases were included in the analysis if there was a pathological diagnosis of pancreatic adenocarcinoma and the patient had undergone pancreaticoduodenectomy between the years 1999 and 2009. Prior to resection, all patients were reviewed at multidisciplinary tumor conference with a pancreatic-hepatobiliary surgeon present and offered resection when deemed appropriate. Metastatic disease and local invasion of adjacent structures precluded resection. Patients with prohibitive medical comorbidities were not resected. Endoscopic retrograde cholangiopancreatography (ERCP) with biliary stent insertion was performed preoperatively if the patient was jaundiced. In majority of patients, endoscopic or computed tomography– guided biopsy was performed to obtain tissue diagnosis. Patients considered to have resectable tumors underwent pancreaticoduodenectomy with lymphadenectomy. Portal vein resection and reconstruction was performed when necessary. Bile duct and pancreatic margins were routinely submitted as frozen sections and attempts at additional resection were made when margins were positive. A negative

Results Of the patients in the in the pancreatico-hepatobiliary database, 365 had a diagnosis of pancreatic adenocarcinoma at the head of the pancreas and underwent pancreaticoduodenectomy with curative intent. There were 182 females (49.8%) and 183 males (50.2%). The median age was 66 years and the median body mass index (BMI) was 25.6. One hundred forty-one patients (38.5%) reported a history of diabetes mellitus. Two hundred thirty-one patients (63.1%) had a preoperative ERCP with stent placement. The median and mode American Society of Anesthesiologists (ASA) classes were both 3. Eighty-seven patients (23.8%) had previous cholecystectomy before diagnosis of pancreatic cancer. Thirty-one patients (35.6%) had recent cholecystectomy (within past 2 years) and 44 patients (50.6%) had remote cholecystectomy (⬎2 years prior). Twelve patients (13.8%) underwent cholecystectomy at an unknown time before pancreaticoduodenectomy. Median age (68 years vs 65 years, P ⫽ .48), BMI (25 vs 25.6, P ⫽ .8), diabetes (33.3% vs 55.4%, P ⫽ .06), ASA class (P ⫽ .22), and preoperative stent placement (70.1% vs 61.2%, P ⫽ .13) were comparable in patients with previous cholecystectomy compared with those without. There was a significant difference in gender (33.3% vs 55.4% male gen-

S.C. Schiffman et al.

Prior cholecystectomy and pancreatic adenocarcinoma

der, P ⫽ .003) between the groups. Operative time (240 vs 240 minutes, P ⫽ .76), estimated blood loss (400 vs 500 mL, P ⫽ .24), intraoperative transfusion (47.2% vs 48.2%, P ⫽ .91), portal vein resection (19.5% vs 14.4%, P ⫽ .25), length of stay (10 vs 10 days, P ⫽ .09), and adjuvant therapy (48.3% vs 58.6%, P ⫽ .2) were not statistically different between patients with prior cholecystectomy compared with those without (Table 1). Tumor pathology did not show any statistical difference between patients with prior cholecystectomy versus patients without. Tumor size (3 vs 3 cm, P ⫽ .89), differentiation (P ⫽ .67), angiolymphatic invasion (43.7% vs 41.7%, P ⫽ .69), perineural invasion (85.1% vs 80.2%, P ⫽ 54), and nodal metastasis (65.5% vs 61%, P ⫽ .43) were not significantly different. Patients with prior cholecystectomy had lower rate of positive margin resection than those without (16.2% vs 31.5%, P ⫽ .003). Postoperatively, there was a statistical difference between prior cholecystectomy and no cholecystectomy in 60-day mortality rate (8.1% vs 2.9%, P ⫽ .034). There was no difference in complication rate (71.3% vs 69%, P ⫽ .75), presence of major complication (26.4% vs 25.9%, P ⫽ .9), infection rate (39.1%, 33.8%, P ⫽ .37), and 30-day mortality (2.3% vs 1.4%, P ⫽ .58) (Table 1).

Table 1 Comparison of factors in patients with previous cholecystectomy to those without history of cholecystectomy Chole Age BMI Male gender DM ASA class Stent Operative time (min) EBL (mL) Transfusion LOS (d) Adjuvant therapy PV resection Differentiation Well Moderate Poor Unknown Size (cm) Nodal metastases LVI Perineural invasion R1 margin Comp rate Major comps Infections 30-d mortality

No chole

68 25 33.30% 40.50%

65 25.6 55.40% 47.10%

70.10% 240 400 47.20% 10 48.30% 19.50%

61.20% 240 500 48.20% 10 58.60% 14.40%

5.75% 62.07% 29.89% 2.30% 3 65.50% 43.70% 85.10% 16.20% 71.30% 26.40% 39.10% 2.30%

5.76% 62.59% 29.50% 2.16% 3 61% 41.70% 80.20% 31.50% 69% 25.90% 33.80% 1.4

P value .48 .8 .003 .062 .228 .13 .763 .236 .908 .09 .2 .249 .66

.89 .43 .69 .54 .003 .745 .9 .37 .58

ASA ⫽ American Society of Anesthesiology classification system; BMI ⫽ body mass index; Chole ⫽ cholecystectomy; Comp ⫽ compilcations; DM ⫽ diabetes mellitus; EBL ⫽ estimated blood loss; LOS ⫽ length of stay; PV ⫽ portal vein; LVI ⫽ angiolymphatic/lymphovascular invasion.

521

Table 2 Complications stratified by cholecystectomy versus no cholecystectomy Grade

Chole (n)

Chole

No chole (n)

No chole

1 2 3 4a 4b 5 Total

11 22 13 4 2 4 56

19.64% 39.29% 23.21% 7.14% 3.57% 7.14%

36 88 48 9 6 9 196

18.37% 44.90% 24.49% 4.59% 3.06% 4.59%

Chole ⫽ cholestectomy.

Overall, 252 patients (69%) sustained postoperative complications. Of these, 95 patients (37.7%) had a major complication (grade 3 or higher). There was no difference in complication rate or grade between patients with prior cholecystectomy and those without (Table 2). There were 31 patients (35.6%) who underwent recent (within 2 years) cholecystectomy. When these patients were compared with patients without previous cholecystectomy, there was no significant difference in epidemiological factors, tumor pathological factors, morbidity, or overall survival. The overall median and mean survival for all patients (n ⫽ 365) were 16 months and 25.7 months, respectively (Fig. 1). Median and mean survival were 14 months and 22.2 months, respectively, for patients with history of cholecystectomy and 16 months and 26.2 months, respectively, for those without (P ⫽ .25) (Fig. 2). Operative mortality was 1.6%. The median follow-up time was 11 months. At the end of the follow-up period, 239 patients (65.5%) had died, and of these, 205 patients (85.8%) were classified as DOD (died of disease) and 34 patients (14.2%) were DOC (died of other causes). Of the 126 survivors (34.5%), 90 patients (71.4%) were classified as NED (no evidence of disease) and 36 patients (28.6%) were AWD (alive with disease). On univariate Cox regression analysis, tumor size (P ⫽ .0023), perineural invasion (P ⫽ .003), angiolymphatic invasion (P ⫽ .014), nodal metastasis (P ⬍ .001), and positive margins (P ⬍ .001) were statistically significant predictors of survival. Tumor size and angiolymphatic invasion remained significant on multivariate analysis by Cox proportional hazards model. The hazards ratio, 95% confidence interval, and probability values are shown in Table 3.

Comments There was no statistical difference in overall survival in patients with pancreatic cancer with a history of cholecystectomy versus those without history of cholecystectomy. There was no difference in most clinical and epidemiological factors in the patients who underwent prior chole-

522

The American Journal of Surgery, Vol 201, No 4, April 2011

Figure 1

Overall survival of 365 patients status post pancreaticoduodenectomy for pancreatic adenocarcinoma.

cystectomy compared with those who had not. We did find a statistically significant difference in gender distribution (P ⫽ .003). There was a much larger percentage of female patients in the previous cholecystectomy group (66.6% vs 44.6%). This is not surprising given the increased prevalence of gallstones and higher rate of cholecystectomy in female patients.17 Obesity, which is another risk factor for gallstones,17,18 was not significantly different between the groups in this study. There was no difference in tumor pathology, including nodal metastasis, size, differentiation, perineural invasion, and angiolymphatic invasion between patients with prior

Figure 2

cholecystectomy and those without. Patients with previous cholecystectomy had significantly lower rate of R1 resection compared with those without prior cholecystectomy (16.2% vs 31.5%, P ⫽ .003). Additionally, when we stratified our cholecystectomies into recent (within 2 years) and remote, we did not see any statistical differences in epidemiology, tumor pathology, or overall survival. Prior cholecystectomy was not a significant predictor of survival on regression analysis. In multiple other analyses, EBL was a predictor of survival.4,9 We attempt to minimize intraoperative blood loss and do not routinely transfuse patients unless they have underlying hemodynamic prob-

Kaplan-Meier survival curve for cholecystectomy versus no cholecystectomy.

S.C. Schiffman et al. Table 3

Prior cholecystectomy and pancreatic adenocarcinoma

Predictors of survival Univariate P value

Size Perineural Invasion LVI Nodal Mets Margins

.023 .003 .014 ⬍.001 ⬍.001

Hazard ratio

95% CI

Multivariate P value

1.6

1.2–2.3

.005

0.72 1.42 0.93 1.1

0.5–1.1 1.0–2.0 0.7–1.3 0.7–1.7

.15 .05 .7 .6

CI ⫽ confidence interval; Mets ⫽ metastases; LVI ⫽ angiolymphatic/lymphovascular invasion.

lem. Additionally, tumor size (⬍3 cm),1,4,9,19,20 negative lymph node status,4,9 –11,19 –21 perineural invasion, angiolymphatic invasion, and negative margins4,19,21 predicted improved survival, which was consistent with multiple past studies. In our study, differentiation was not a statistically significant predictor of survival. This was likely due to the small number of well differentiated tumors (5.75%) in our study. Previous studies reported association between pancreatic cancer and cholecystectomy but did not analyze time period between operation and cancer diagnosis.22,23 Thus, the perceived association could actually be due to early cancer symptoms mistaken for biliary disease.13,24 Ekbom et al25 found modest excess risk of pancreatic cancer following cholecystectomy. The standardized incident ratio was 1.52 at 4 years and 1.35 at 15 years. However, the standardized incident ratio was 16.6 at 1 year, which suggests that early symptoms of pancreatic cancer may have been misdiagnosed as biliary disease and led to cholecystectomy. Multiple studies have found significant association between pancreatic cancer and recent (within 1 year) cholecystectomy.12,23,25 The association disappeared after patients who underwent recent cholecystectomy were excluded from analyses. This further suggests that symptoms of pancreatic cancer may be inappropriately attributed to biliary disease. Ghadimi et al13 all reported a 9% incidence of cholecystectomies within 2 years before diagnosis of pancreatic cancer. They attribute this to misdiagnosis of atypical biliary symptoms. Also, the study asserts that both laparoscopic and open cholecystectomy allow for misdiagnosis. In animals, cholecystectomy has been shown to induce pancreatic hyperplasia and hypertrophy. The mechanism is unclear, but may be attributed to an increase in cholecystokinin, which has trophic effects on the pancreas.24 The relationship between cholecystokinin and pancreatic adenocarcinoma remains unclear. Further studies are needed to elucidate the relationship between pancreatic cancer and cholecystectomy. The causative role of cholelithiasis, cholecystectomies and other biliary disorders with pancreatic cancer remains unclear.24 In our study, we attempted to analyze whether this delay in recognizing the diagnosis of pancreatic cancer adversely af-

523

fected overall survival. We did not find a significant survival difference between patients who underwent previous cholecystectomy (recent or remote) and those who did not. Furthermore, we did not find any difference in tumor pathology, morbidity or mortality in patients with recent cholecystectomy compared with those without prior cholecystectomy. A limitation of this study was that we only analyzed patients that were amenable to pancreaticoduodenectomy resection. Perhaps, some patients status post recent cholecystectomy were not amenable to curative resection secondary to local invasion or metastatic disease and thus were excluded from our analysis. This is much evidence that early detection of pancreatic adenocarcinoma offers the best chance for curative resection and long-term survival. Further improvements in survival will likely come from efforts to achieve early diagnosis and improvements in adjuvant therapy. Clinicians should maintain a low threshold for investigating biliary symptoms that persist after cholecystectomy to facilitate early diagnosis and treatment of pancreatic cancer.

References 1. Cleary SP, Gryfe R, Guindi M, et al. Prognostic factors in resected pancreatic adenocarcinoma: analysis of actual 5-year survivors. 2004; 198:722–31. 2. Stojadinovic A, Brooks A, Hoos A, et al. An evidence-based approach to the surgical management of resectable pancreatic adenocarcinoma. 2003;196:954 – 64. 3. Vento P, Mustonen H, Joensuu T, et al. Impact of preoperative chemoradiotherapy on survival in patients with resectable pancreatic cancer. 2007;13:2945–51. 4. Sohn TA, Yeo CJ, Cameron JL, et al. Resected adenocarcinoma of the pancreas—six hundred and sixteen patients: results, outcomes, and prognostic indicators. 2000;4:567–79. 5. Cameron JL, Pitt HA, Yeo CJ, et al. One hundred and forty-five consecutive pancreaticoduodenectomies without mortality. 1993;217: 430 –5. 6. Conlon KC, Klimstra DS, Brennan MF. Long-term survival after curative resection for pancreatic ductal adenocarcinoma. Clinicopathologic analysis of 5-year survivors. 1996;223:273–9. 7. Gouma DJ, van Geenen RC, van Gulik TM, et al. Rates of complications and death after pancreaticoduodenectomy: risk factors and the impact of hospital volume. 2000;232:786 –95. 8. Martin RC, Brennan MF, Jaques DP. Quality of complication reporting in the surgical literature. 2002;235:803–13. 9. Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas. Two hundred and one patients. 1995;221:721–31. 10. Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. 1997;226:248 –57. 11. Yeo CJ, Sohn TA, Cameron JL, et al. Periampullary adenocarcinoma: analysis of 5-year survivors. 1998;227:821–31. 12. Gullo L. Risk of pancreatic and periampullary cancer following cholecystectomy. 1999;10 Suppl 4:127– 8. 13. Ghadimi BM, Horstmann O, Jacobsen K, et al. Delay of diagnosis in pancreatic cancer due to suspected symptomatic cholelithiasis. 2002; 37:1437–9. 14. Malouf AJ, Murray AW, MacGregor AB. Major intra-abdominal pathology missed at laparoscopic cholecystectomy. 2000;87:1434 –5.

524 15. Siriwardena A, Samarji WN. Cutaneous tumour seeding from a previously undiagnosed pancreatic carcinoma after laparoscopic cholecystectomy. 1993;75:199 –200. 16. Slim K, Pezet D, Clark E, et al. Malignant tumors missed at laparoscopic cholecystectomy. 1996;171:364 –5. 17. Friedman GD. Natural history of asymptomatic and symptomatic gallstones. 1993;165:399 – 404. 18. Salinas G, Velasquez C, Saavedra L, et al. Prevalence and risk factors for gallstone disease. 2004;14:250 –3. 19. Benassai G, Mastrorilli M, Quarto G, et al. Factors influencing survival after resection for ductal adenocarcinoma of the head of the pancreas. 2000;73:212– 8. 20. Geer RJ, Brennan MF. Prognostic indicators for survival after resection of pancreatic adenocarcinoma. 1993;165:68 –72.

The American Journal of Surgery, Vol 201, No 4, April 2011 21. Katz MH, Wang H, Fleming JB, et al. Long-term survival after multidisciplinary management of resected pancreatic adenocarcinoma. 2009;16:836 – 47. 22. Cuzick J, Babiker AG. Pancreatic cancer, alcohol, diabetes mellitus and gall-bladder disease. 1989;43:415–21. 23. Hyvarinen H, Partanen S. Association of cholecystectomy with abdominal cancers. 1987;34:280 – 4. 24. Chow WH, Johansen C, Gridley G, et al. Gallstones, cholecystectomy and risk of cancers of the liver, biliary tract and pancreas. 1999;79: 640 – 4. 25. Ekbom A, Yuen J, Karlsson BM, et al. Risk of pancreatic and periampullar cancer following cholecystectomy: a population-based cohort study. 1996;41:387–91.