Role of splenectomy in gastric cancer surgery: adverse effect of elective splenectomy on longterm survival

Role of Splenectomy in Gastric Cancer Surgery: Adverse Effect of Elective Splenectomy on L o n g t e r m Survival Harold J. Wanebo, MD, FACS,B.j. Kennedy, MD, FACS,David P. Winchester, MD, FACS, Andrew K. Stewart, MA, and Amy M. Fremgen, PhD Background: Splenectomy, and in some cases pancreafico splenectomy, has been advocated by surgeons in an effort to improve clearance of metastatic nodes to splenic hilum (node 10) and splenic artery (node 11). Although splenectomy has known effects on increasing morbidity and even mortality after a variety of surgical maneuvers including gastrectomy, the longterm effect on survival is controversial. The purpose of this study is to review and analyze the effect o f splenectomy on survival in patients having curative gastrectomy for stomach cancer. /

Methods: We reviewed the role of splenectomy in patients having curative gastrectomy in a data base of stomach cancer patients that had been collected in 1987 as part of an American College of Surgeons Patterns of Care Study. This analysis had involved 18,344 patients, of whom 11,252 were first diagnosed in 1982 as part of a longterm study, and 7,092 were first diagnosed in 1987 as part of a shortterm study. From the two data collection periods information was available on 12,439 patients who received cancer directed abdominal surgery; 21.2% of these patients received a splenectomy. Among the 3,477 patients reported as having a curative gastrectomy (pathologically clear margins), 26.2% received a splenectomy. Results: The operative mortality was 9.8% with splenectomy and 8.6% without splenectomy. In patients having a curative gastrectomy, the 5-year observed survival rate was 20.9% in patients having splenectomy versus 31% in patients who did not receive splenectomy (p < 0.0001). Examination of differences in survival by stage of diagnosis showed significantly reduced survival outcomes among patients with stage II and III, but not for those diagnosed with stage I or 1V disease. The pattern of recurrence was moderately different with a larger proportion of patients having distant metastases among the group of patients who had undergone splenectomy Received October 28, 1996; Revised April 3, 1997; Accepted April 3, 1997. From the Roger Williams Medical Center, Providence, RI. Correspondence address: Harold J. Wanebo, MD, Professor of Surgery, Director of Surgical Oncology, Brown University, Chief of Surgery, Roger Williams Medical Center, 825 Chalkstone Avenue, Providence, RI 02908. © 1997 by the American College of Surgeons Published by Elsevier Science Inc.

177

compared with the patients who had not, 29% and 15.5%, respectively. Whether these differences are in~ herent in the splenectomy or in the associated cofactors was not determined in this study. Conclusions: The data suggest elective splenectomy should generally be avoided in patients with stage II and III gastric cancer. In patients with resectable proximal advanced (stage IV) cancer or who have extension to spleen and pancreas or macroscopic nodal metastases to splenic hilum, splenectomy might be necessary to facilitate complete removal of the tumor in an effort to achieve longterm tumor control. The importance of surgical judgment is emphasized as the major deciding factor in determining the need for splenectomy in the individual cancer patient. (J Am Coil Surg 1997;185: 177-184. © 1997 by the American College of Surgeons) T h e role o f splenectomy in the surgical managem e n t of gastric c a n c e r is controversial. The concept o f e x t e n d e d lymph n o d e dissection a n d its c o n t r i b u t i o n to l o n g t e r m control a n d survival is accepted by m a n y surgical centers, in particular, those in Japan, Korea, Taiwan, a n d parts of Europe (1-4). D e t e r m i n a t i o n o f its therapeutic efficacy is awaiting the o u t c o m e o f o n g o i n g r a n d o m ized clinical trials, b u t there is c o n c e r n a b o u t the s h o r t t e r m adverse effects on morbidity a n d mortality (5, 6). T h e use o f splenectomy to facilitate removal o f lymph n o d e s in the splenic h i l u m a n d along the splenic artery underlies the rationale for the procedure. Soga a n d colleagues are a m o n g the authors who have e m p h a s i z e d the n e e d for splenectomy to facilitate removal o f nodes in the stomach or along the splenic artery (1). T h e true yield o f n o d a l metastases at the splenic h i l u m is d e p e n d e n t on the a n a t o m i c subsite of the primary cancer, its d e p t h of invasion (thickness), a n d w h e t h e r there is p e n e t r a t i o n t h r o u g h the serosa. According to the J a p a n e s e studies, the incidence o f metastases to splenic hilar n o d e s ranges f r o m 0-1.9% for the m i d d l e third a n d distal third gastric cancers, to 15.5% for proximal third gastric cancer, a n d 20.7% in patients with cancerous inISSN 1072-7515/97/$17.00 PII S1072-7515(97)00047-1

178 J AM COLL SURG AUGUST1997

VOLUME185:177-184

vasion of the total stomach (7). Penetration by the primary t u m o r in or t h r o u g h the serosa can increase the n o t e d metastatic rate (8). Many authors do n o t support p e r f o r m i n g a splenectomy because of its proved adverse effect on morbidity a n d postoperative mortality and its suspected effects on t u m o r growth and survival (9-12). O u r study is directed at reviewing effects of splenectomy on the shortterm and longterm o u t c o m e a m o n g patients who have u n d e r g o n e curative gastrectomy. Me~o~

The details of the study have been previously rep o r t e d (13). A patient care evaluation study of cancer of the stomach was initiated by the American College of Surgeons in 1987 using data collected from m o r e than 700 t u m o r registries t h r o u g h o u t the United States. Data on patients diagnosed in 1982 f o r m e d a longterm study, a n d patient data collected in 1987 represented a shortt e r m study. Each t u m o r registry was asked to submit data on 25 patients for each of the study periods. Primary gastric cancer consisted of all sites listed in the International Classification of Disease in Oncology c o d e d from 151.0 to 151.9 and included esophagogastric j u n c t i o n tumors and all adenocarcinomas of various types. Excluded were leiomyosarcoma, leioblastoma, lymp h o m a , and squamous cancer. A detailed questionnaire recorded data on information on patient demographics, disease characteristics, and patterns of treatment. This included the anatomic location of the tumor, its size, histology, a n d histologic characteristics such as type, grade, a n d Lauren's classification (14). Tumors were staged according to the American J o i n t Commission on Cancer (AJCC) staging system (15). Specific details about the initial a n d subseq u e n t therapy by surgery, radiation, and chemotherapy were requested. T h e various surgical procedures and types of gastric resection c o m m o n l y used were listed. Removal of tissue beyond the stomach was labeled as e x t e n d e d gastrectomy. T h e use of adjuvant therapy after surgery and the overall patterns of recurrence and its t r e a t m e n t were recorded. T h e extent of resection beyond the stomach was n o t e d and categorized as resected (malignant or nonmalignant) or n o n r e s e c t e d (malignant, nonmalignant, or u n k n o w n condition). The resection of extra gastric organs such as colon, esophagus, gallbladder, o m e n t u m , pancreas, or spleen, a n d the extent of lymph n o d e dissection were noted. Information was requested regarding n o d e dissections of the following sites:

perigastric, gastric, c o m m o n hepatic, celiac, splenic; or other. The focus on the study was the shortterm, postoperative survival and l o n g t e r m survival, reflecting possible effects of the disease. T h e latter focus was based on an analysis of data in patients in the longterm study (1982) who h a d a curative resection; all disease was resected with pathologically clear margins in these patients. All data were analyzed using a standard statistical software package (16, 17). Survival curves reflecting observed deaths from all causes were generated using the actuarial life table m e t h o d . Differences between survival curves were evaluated using the Wilcoxon statistic, a n o n p a r a m e t r i c statistic in which significance tests are two-sided a n d unadjusted for multiple comparisons (18-20). Results

T h e initial r e p o r t provided data on 18,344 patients with stomach cancer of w h o m 11,252 were in a longterm study group (1982) and 7,092 were in the shortterm study group (1987). C o m b i n i n g data from both study periods, 12,434 patients underwent an abdominal surgical procedure, of these 7,770 patients h a d a gastrectomy. Curative resection was p e r f o r m e d in 3,661 patients of w h o m 912 (26.2%) had a splenectomy a n d 2,565 did not. The splenectomy status was u n k n o w n in 184 patients. Of those patients having splenectomy, 10.3% had cancer in the spleen. A m o n g patients receiving a gastrectomy with pathologically clear margins, there was a difference in the frequency of splenectomies d e p e n d i n g on the anatomic subsite of the tumor. T h e p r o p o r t i o n of patients receiving a splenectomy ranged from 41% in patients having an u p p e r third gastric resection, to 22.8% for middle third lesions a n d 10.1% for distal third lesions. A m o n g patients with total stomach involvement over half (54.3%) h a d their spleens resected. Mortality a m o n g patients after gastrectomy with or without splenectomy was similar. T h e overall mortality was 9.8% in the gastrectomy patients who u n d e r w e n t splenectomy and 8.6% in those who did not. The increase occurred in all the major anatomic gastric sites. T h e effect of splenectomy on survival outcomes varied by resection site. Observed 5-year survival rates are c o m p a r e d according to site of t u m o r location and resection in Table 1. In each case, patients u n d e r g o i n g splenectomy experienced p o o r e r survival outcomes, ranging 4 - 9 percentage points, c o m p a r e d with those who did n o t receive a

W a n e b o et al

ROLE OF S P L E N E C T O M Y IN GASTRIC C A N C E R SURGERY

Table 1. Gastrectomy with Clear Margins: Effect of Splenectomy on Survival Splenectomy Resection

Table 2. Splenectomy in Patients Having Curative Gastrectomy: Survival by Stage

No Splenectomy

5-year observed

AJCC Stage

5-year observed

site n survival(%) Distal 133 27.1 Middle 127 32.2 Proximal 401 17.8 Total stomach 91 10.2

179

n survival(%) 1,022 32.5 424 41.6 584 22.1 75 15.2

p 0.0838 0.2307 0.2231 0.6790

Splenectomy

n

5-year observed survival (%)

yes no yes no yes no yes no

146 599 121 554 378 805 149 357

50.3 56.1 25.4 37.6 13.8 18.1 3.9 4.9

I II III 1V

splenectomy. In every case, with the e x c e p t i o n o f total stomach resection, the survival rates b e t w e e n patients w h o h a d u n d e r g o n e s p l e n e c t o m y a n d those w h o h a d n o t is markedly different. W i t h o u t regard to anatomic subsite o f the t u m o r the observed 5-year survival rate o f patients having curative gastrectomy was 20.9% in the patients with s p l e n e c t o m y and 31% in patients Without (p < 0.0001) (Fig. 1). O b s e r v e d 5-year smwival rates are shown in Table 2. D e p e n d i n g on the stage o f disease there were significant differences in the survival rates b e t w e e n patients w h o received a s p l e n e c t o m y a n d those w h o did n o t (Fig. 2A). This was particularly clear in the calculated 5-year survival rates of patients d i a g n o s e d with stage 1I a n d III disease (Figs. 2B, C). T h e n u m e r i c differences in the 5-year survival rates for stage I patients was suggestive o f a significant difference; however, there was n o significant difference a m o n g stage 1V patients be-

Median survival time (months) 60.6 75.5 25.3 35.6 15.2 20.0 9.7 8.8

p 0.0726 0.0230 0.0030 0.2445

tween those who did and those w h o did n o t receive a splenectomy. Data were available on the effect of s p l e n e c t o m y on the pattern o f r e c u r r e n c e for 660 patients. T h e r e were 155 patients in the g r o u p o f patients with s p l e n e c t o m y a n d 505 patients w h o did n o t have splenectomy. T h e r e were no noticeable differences in local or regional r e c u r r e n c e in the patients having a s p l e n e c t o m y c o m p a r e d t o the patients n o t having a splenectomy. T h e local a n d regional r e c u r r e n c e rates were 7.7% a n d 9.1%, respectively, in the s p l e n e c t o m y patients comp a r e d to 6.2% a n d 5.1%, respectively, in the nons p l e n e c t o m y patients. T h e p r o p o r t i o n o f patients who were• never disease free was 25.9% in spl,enectomy patmnts a n d 23.9% in those without splenectomy. In contrast, 26.6% of the patients with splen e c t o m y r e m a i n e d disease free while 47.7% o f the g r o u p without s p l e n e c t o m y were r e p o r t e d as hay-

Overall Survival 100

% Cumulative Survival

80

60

40

20

No Splenectomy .e.Splenectomy P = 0.0001 0

No Splenectomy Splenectomy

100 100

68.7 61.1

50.1 39

40.1 28.6

4 34.6 23.7

5 31 20.9

Years from Diagnosis

FIG 1. Overall survival of 1,982 stomach cancer patients receiving curative gastrectomy, with known American Joint Commission on Cancer (AJCC) stage of disease at diagnosis. Splenectomy, n = 908; No Splenectomy, n = 2,555.

180

J AM COLL SURG AUGUST 1997

VOLUME 185:177--184

AJCC Stage I lOO

% Cumulative Survival

80

60

40

20

I N o Splenect°m'I~ p = 0.73 •e, Splenectomy ] o No Splenectorny Splenectomy

o 100 100

1 84.9 80.1

2 75.8

64.2

3 67.9 58.7

4 61.1 54.5

5 56.1 50.3

Years from Diagnosis

A

AJCC Stage II % Cumulative Survival 10o

80

60

40

20 i , No Splenectomy •e-Splenectomy P = 0.023

No Splenectomy

Splenectomy

100 100

58

77.6 77

51.3

49.7 35.8

4

5

43.5 29.7

37.6 25.4

Years from Diagnosis

B FIG 2. Stage-specific survival of 1,982 stomach cancer patients receiving curative gastrectomy, with known American Joint Commission on Cancer (AJCC) stage of disease at diagnosis. (A) Stage I: Splenectomy, n = 146; No Splenectomy, n = 599. (B) Stage II: Splenectomy, n = 170; No Splenectomy, n = 554. (C) Stage III: Splenectomy, n = 378; No Splenectomy, n = 805. (D) Stage IV: Splenectomy, n = 149; No Splenectomy, n = 357. (Figure continues.)

ing remained disease flee. The proportion of patients r e p o r t e d w i t h d i s t a n t metastasis as t h e first m a n i f e s t a t i o n o f d i s e a s e r e c u r r e n c e was m a r k e d l y different between the group receiving and the group not receiving splenectomy. Among patients who received a s p l e n e c t o m y 31.5% were r e p o r t e d w i t h d i s t a n t m e t a s t a t i c r e c u r r e n c e a n d 17.1% o f

patients without splenectomy experienced distant m e t a s t a t i c r e c u r r e n c e , p < 0.001.

Discussion O p e r a t i o n f o r gastric c a n c e r in J a p a n a n d in Western countries has been frequently combined with

Wanebo et al

ROLE OF SPLENECTOMY IN GASTRIC GANCER SURGERY

181

AJCC Stage III 100

% Cumulative Survival

o

60

40

2O •e, No Splenectomy i , Splenectomy

m

P = 0.003

0 I No Splenectomy Splenectomy

0 100 100

1 67 56.2

2 41.6 32.9

3 26.6 21.3

4 20.4 15.9

5 18.1 13.8

Years from Diagnosis

C

AJCC Stage IV 100

% Cumulative Survival

80

60

40

20

:N m

No Splenectomy Splenectomy

0 100 100

,1 31.7 38.1

12.5 15.4

7 7.7

Years from

4 5.6 6.2

5 4.9 3.9

Diagnosis

D FIG. 2. (Continued) resection of the pancreatic body and tail or of the spleen to facilitate a curative resection. The therapeutic n e e d for distal pancreatectomy and splen e c t o m y in the presence of direct invasion is obvious. T h e elective use of the pancreatectomy or splenectomy in the absence of direct invasion is m o r e problematic. The m a n a g e m e n t of potential lymph n o d e metastases to the nodes in the splenic h i l u m (14), and those along the splenic artery (15) have been a major rationale for elective sple-

nectomy or pancreatic splenectomy. Many surgeons have considered it essential to resect the spleen to allow complete removal of the hilar nodes (14) and the perisplenic nodes (15). In 1967 Kanai q u o t e d by Okajima and Isozaki (7) examined serial sections of the pancreatic body, tail, and spleen resected from patients in w h o m nodes 10 a n d 11 were considered dissected. He d e m o n s t r a t e d that 74% of nodes 10 and 11, and 65% of the nodes involved with metastases

182 J AM COLL SURG AUGUST19.97 VOLUME185:177--184 T a b l e 3. R e p o r t e d S t u d i e s o f S p l e n e c t o m y D u r i n g G a s t r e c t o m y f o r S t o m a c h C a n c e r Author (reference n u m b e r )

Patients

Splenectomy

n

Morbidity (%)

Mortality (%)

Survival (%)

Brady (26)

392

Stipa (22)

646

Adachi (24)

116

Griffith (23)

195

Otsuji (25)

245

yes no yes no yes no yes no yes no

163 229 8I 565 58 58 76 119 154 91

45 21 na na 58 26 41 14 57 37

5 3 na na na na 12 na na na

38* 50 31" 42 261 47 45* na 46* 47

*5-year survival. tl0-year survival. na, not available.

were left undissected, emphasizing the n e e d for pancreatic splenectomy to obtain complete dissection of nodes 10 and 11. In contrast, Maruyama and colleagues have r e p o r t e d a complete dissection of nodes 10 and 11 during his pancreaspreserving operation (21), which included removal of the splenic artery and vein with preservation of spleen and pancreas. Complete removal was d e m o n s t r a t e d by lymphatic flow studies in histologic examination, and they reported an excellent prognosis after this pancreaspreserving operation (21). The n e e d to p e r f o r m a splenectomy to facilitate removal of the splenic hilar nodes a n d nodes along the splenic artery is highly influenced by the site of the primary cancer. In a review of the Japanese experience with splenectomy by Okajima and Isozaki, the incidence of hilar nodal metastases was stated at 15.5% for u p p e r third cancers, 1.9% f o r middle third cancers, a n d 0% for lower third cancers, b u t this figure was 20.7% when the entire stomach was involved. These n u m b e r s are also influenced by the extent of the primary cancer, (ie, serosal penetration is associated with a higher metastatic rate). T h e location and stage of the t u m o r has a major effect on the n e e d for splenectomy to allow adequate hilar control. T h e concept of an effective elective splenectomy is governed by the n u m e r o u s associated tum o r factors. In a review by Stipa of 646 p a t i e n t s u n d e r g o i n g curative gastrectomy for cancer of the stomach in a R o m e hospital, splenectomy was associated with significantly impaired survival in univariate analysis (22). It shared differences with t u m o r location, gastrectomy, histologic type (diffused versus intestinal), T size, and nodal status. In a multivariate analysis of nine clinical, pathologic, or treatment-related variables, the only independ e n t predictors of death included cirrhosis, distant lymph n o d e metastasis, and total gastrectomy

with loss of significance ofsplenectomy and lymphadenectomy. In this study, separate analyses of the patient subsets revealed that radical lymphadenectomy, removal of N2 nodes, h a d survival benefit for patients with T1, T2, and stages 0 to I disease (Table 3). In the study by Griffith and colleagues of 195 curative resections for carcinoma of the stomach, there was a significant increase in operative mortality associated with splenectomy and an increase in complications (23). Five-year survival with gastrectomy and splenectomy was 45% c o m p a r e d with 71% after gastrectomy alone (p < 0.01). A Cox proportional hazards m o d e l showed that splenectomy had an adverse influence on patient survival, and the authors (23) suggested that its routine use in the course of radical resection for carcinoma of the stomach should be abandoned. The influence of splenectomy on survival by stage showed no significant impact of splenectomy in pathologic stage I and II disease, b u t in stage III there was increased survival in patients having gastrectomy without splenectomy (35% versus 25%). W h e n t u m o r factors were analyzed in the Cox model, metastases to nodes, d e p t h of tumor, a n d penetration were i m p o r t a n t predictors of survival and indicators whether to p e r f o r m a splenectomy. The type of gastric resection h a d no influence on survival (23) (Table 3). Adachi and associates evaluated splenectomy in node-positive gastric cancer patients, a group expected to have a high n u m b e r of nodal metastases to the splenic hilum (24). A m o n g patients receiving a curative resection, no significant difference was f o u n d in those having splenectomy versus no splenectomy, whether one examined the total patient population or examined cases with N1 n o d e or N2 n o d e metastasis. The authors did d e m o n strate a higher percentage of complications in patients who had a splenectomy (24) (Table 3).

Wanebo

e t al

ROLE OF SPLENECTOMY IN GASTRIC CANCER SURGERY

Otsuji a n d colleagues examined the role of splen e c t o m y in patients treated at Kyoto University who had u n d e r g o n e total gastrectomy (25). Total gastrectomy with splenectomy was p e r f o r m e d to facilitate dissection of splenic hilar nodes. T h e authors f o u n d no significant difference in the 5year survival of patients u n d e r g o i n g total gastrectomy with or without splenectomy. No correlation was reported for any prognostic factor with the splenectomy. Using logistic regression analysis, sp!enectomy was associated i n d e p e n d e n t l y with postoperative complications. With respect to longterm survival, this analysis indicated that only the presence of residual t u m o r h a d significant impact. T h e authors r e c o m m e n d c o m b i n e d splenectomy in patients Where splenic hilar nodes are involved because of the difficulty of p e r f o r m i n g a complete splenic hilar dissection while still preserving the spleen (see Table 3). In a study from Memorial S10an-Kettering Cancer Center, Brady and colleagues reviewed the records o f 392 patients who u n d e r w e n t curative gastrectomy for adenocarcinoma; there were 12 factors, including splenectomy, associated with p o o r prognosis by univariate analysis (26). Multivariate analysis identified 6 of these factors b u t excluded splenectomy as an i n d e p e n d e n t predictor of death from gastric cancer. T h e apparent adverse effect of splenectomy was believed to be its association with other significant risk factors. Postoperative complications were significantly higher in the splenectomized patients, 45% versus 21%. A high percentage of these were infectious complications. Unfavorable prognostic variables associated with splenectomy included gastrectomy (not distal), T4 primary lesion, stage greater than IA, positive nodes, and a positive resection margin. T h e authors c o n c l u d e d that splenectomy is n o t directly related to survival, but morbidity increases if curative gastrectomy is p e r f o r m e d when the spleen is close to or is invaded by t u m o r (Table 3). O u r study represents an evaluation of the surgical m a n a g e m e n t of gastric cancer in m o r e than 700 hospitals across the United States. The data indicate that 21.1% of all the r e p o r t e d cases of gastrectomy for stomach cancer included splenectomy. Patients with u p p e r third lesions were m o r e likely to have a splenectomy than patients with lower third lesions, 41% and 10.1%, respectively. T h e operative mortality was n o t affected significantly by splenectomy, 9.8% with and 8.6% without. T h e r e were significant differences in survival rates. T h e observed 5-year survival rate for patients who u n d e r w e n t a splenectomy was 20.9% and for

!83

patients without a splenectomy the 5-year survival rate was 31%. Survival rates were significantly impaired by splenectomy in patients with stage II and III, but n o t for patients diagnosed with stage I or IV disease. The p a t t e r n of recurrence in the patients with splenectomy was moderately different from that of the patients without splenectomy; an increased n u m b e r of patients had distant metastases, 29% versus 15.5%, respectively. Whether these differences are i n h e r e n t in the splenectomy per se or in the associated cofactors was n o t d e t e r m i n e d in this study. The putative mechanisms for possible longterm effect of splenectomy on survival in cancer patients are not known with any certainty. Splenectomized patients have diminished antibody responses to particulate antigens, are deficient in tuftsin and phagocytosis p r o m o t i n g peptide, and have decreased levels of IgM p r o p e r d i n (27). Such patients are known to be at risk for septicemia from encapsulated organisms (Streptococcus pneumoniae, Neisseria meningitidis, or Haemophilus influenzae), although they are n o t at increased risk for most nonbacterial infections (28). The influence of splenectomy on longterm immunity is n o t clear (29). T h e surgical decision whether or not to perform a splenectomy is heavily weighted by the numerous t u m o r and patient factors at the time of the operation. Factors likely to influence the n e e d for splenectomy include invasion of the distal pancreas or spleen or the appearance of macroscopic nodal disease, especially in patients with proximal third gastric lesions or in patients with high t u m o r stage who are likely to have nodal metastases. Splenic preservation should be p u r s u e d in patients with early stage cancers, certainly in those with middle-third or distal-third lesions. Although elective splenectomy should be avoided, a surgeon's j u d g m e n t based on the adequacy of the resection and the patient's t u m o r factors remains the most i m p o r t a n t factor d e t e r m i n i n g the n e e d for splenectomy in an individual patient. Conclusions

The effect of splenectomy on shortterm and longterm survival was reviewed in patients with stomach cancer, based on the American College of Surgeons Pattern of Care Study of 1982 and 1987 cases. Splenectomy was p e r f o r m e d in 21.1% of the 12,434 cases r e p o r t e d from both study periods. A m o n g the 3,661 patients who had a curative gastrectomy 26.2% also received a splenectomy. The p r o p o r t i o n of patients u n d e r g o i n g a splenectomy

184 J AM COLL SURG AUGUST1997

VOLUME185:177--184

varied by the anatomic subsite of the tumor, ranging from 10.1% for distal third to 41% for u p p e r third cancer and 54.3% for total gastric involvement. The operative mortality was 9.8% with splenectomy and 8.6% without splenectomy. Splenectomy was associated with impaired 5-year survival rate in patients who h a d a curative gastrectomy and who had pathologically clear margins. Observed 5-year survival rates were 20.9% in patients with splenectomy and 31% in patients without splenectomy (p < 0.0001). Survival differences by stage of diagnosis were significant in patients with stage II or III disease but n o t in patients with stage I or IV disease. T h e patterns of recurrence in patients with splenectomy were moderately different, with increased n u m b e r s having distant metastases, 31.5% c o m p a r e d with 17.1% in patients without sPlenectomy. T h e data suggest that elective splenectomy should be avoided, especially in stages II and III, because it may incur a d d e d risk to longterm survival in patients having curative gastrectomy for stomach cancer. Splenectomy or pancreaticosplenectomy may be required for local tum o r control in selected high-risk patients, those with proximal third cancers who have macroscopic splenic hilar n o d e involvement or direct t u m o r extension. Surgical j u d g m e n t is still the major deciding factor d e t e r m i n i n g the n e e d for splenectomy in the individual cancer patient. References 1. SogaJ, Kobayashi K, Saito J, et al. The role of lymphadenectomy in curative surgery for gastric cancer. World J Surg 1979;3:701-8. 2. Maryuyama K, Okabayashi K, and Kinoshita T. Progress in gastric cancer surgery and its limits of radicality. WorldJ Surg 1987;11:418-25. 3. KimJ, Kwoni OJ, Oh ST, and Ynag HI~ Results of surgery on 6,539 gastric cancer, patients and immunochemo surgery as the best treatment of advanced gastric cancer. Ann Surg 1992;216:269-79. 4. Roder JD, Bottcher K, Siewert R, et al. Prognostic factors in gastric carcinoma: results of German gastric carcinoma study 1992. Cancer 1993;72:2089-97. 5. CushieriA, Fayers B, and FieldingJ. Post operative morbidity after D 1 and D2 resections for gastric cancer: preliminary results of MRC randomized controlled surgical trial. Lancet 1996;347:995-9. 6. Bowenkamp JJ, Songun J, Hermans J, et al. Randomized comparison of morbidity after D 1 and D2 dissection for gastric cancer in 996 Dutch patients. Lancet 1995;345:745-8. 7. Okajima K, and lsozaki H. Splenectomy for treatment of gastric cancer: Japanese experience. World J Surg 1995;19: 537-40.

8. Maruyama K, Gunven P, Okalayashi K, Sasaho M, and Kinoshita T. Lymph node metastases of gastric cancer: general pattern in 1931 patients. Ann Surg 1989;210:596-602. 9. Orita K, Konaga E, Okada T, Kunisada K, Yumura M, and Tanaka S. Effect of splenectomy in tumor-bearing mice and gastric cancer patients. Gann (Tokyo) 1972;68:731-6. 10. Sugimachi K, Kodama Y, Kumashiro R, Kanematsu T, Noda S, and Inokuchi K. Critical evaluation of prophylactic splenectomy in total gastrectomy for stomach cancer. Gann (Tokyo) 1977;71:704-9. 11. Yoshino K, Haruyama K, Nakamura S, et al. Evaluation of splenectomy for gastric carcinoma. Jpn J Gastroenterol Surg 1979;12:944-9. 12. Suchiro S, Nagasue N, Ogawa Y, Sasaki Y, Hirose S, and Yukawa H. The negative effect of splenectomy on the prognosis of gastric cancer. Am J Surgery 1984;148:645-8. 13. Wanebo HJ, Kennedy BJ, ChmielJ, Steele G, Winchester D, and Osteen R. Cancer of the stomach: a patient care study by the American College of Surgeons. Ann Surg 1993;218:583-92. 14. Lauren P. The two histological main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma: an attempt at a histoclinical classification. Acta Pathol Microbiol Scand 1965;64:31-9. 15. Beahrs OH, and Myers MH. Manual for staging of cancer. Philadelphia: J.B. Lippincott; 1988. 16. SPSS for Windows, Advanced Statistics, Release 6.1. Chicago IL: SPSS Inc.; 1994. 17. Morusis MJ. SPSS 6.t. Guide to Data Analysis. Englewood Cliffs, NJ: Prentice Hall; 1996. 18. Kaplan EL, and Meier P. Nonparametric estimation from incomplete observation. J Am Star Assoc 1958;53:457-81. 19. Peto R, and Pile MC. Conservatism in the approximation_£ ( O - E)2/E in the log-rank test for survival data or tumor incidence data. Biometrics 1973;29:579-84. 20. Cox DR. Regression models and life tables. Intl Stat Assoc 1972;34:187-220. 21. Maruyama K, Kitaoka H, Hirata K, Okabayashi K, and Kikuchi S. Is combined resection necessary for no. 10 and no. 11 lymph nodes dissection in radical gastrectomy? Gastroenterol Surg (Shokakigeka) 1984;7:1509-15. 22. Stipa S, DiGiorgio A, Ferri M, and Botti C. Results of curative gastrectomy for carcinoma. J Am Coll Surg 1994;17:567-72. 23. Griffith JP, Sue-Ling HM, Dixon MF, McMahon MJ, Axon ATR, and Johnston D. Preservation of spleen improves survival after radical surgery for gastric cancer. Gut 1995;36: 684-90. 24. Adachi Y, Kamakura T, Mori M, Maehara Y, and Sugimachi K. Role of lymph node dissection and splenectomy in nodepositive gastric carcinoma. Surgery 1994;116:837-41. 25. Otsuji E, Yamaguchi T, Sawai K, Ohara M, and Takahashi T. End results of simultaneous splenectomy in patients undergoing total gastrectomy for gastric carcinoma. Surgery 1996; 120:40-4. 26. Brady MS, Rogatho A, Dent LL, and Shiu MH. Effect of splenectomy on morbidity and survival following curative gastrectomy for gastric carcinoma. Arch Surg 1991;26:35964. 27. Rose WF. The spleen as filter. N EnglJ Med 1987;317:705-6. 28. Schimpff SC, O'Connell MJ, Greene WH, et al. Infections in 92 splenectomized patients with Hodgkin's disease: a clinical review. Am J Med 1995;59:695-701. 29. Penn I. Immunocompetence and cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Biologic Therapy of Cancer. Philadelphia: J. B. Lippincott; 1995;103-18.