Splenectom y

Between February 1980 and December 1985, splenectomies were performed in 13 patients with advanced (Stage 111 and IV) chronic lymphocytic leukemia (CLL). Patients had advanced disease for a median of 32 months at the time of splenectomy and all were refractory to chemotherapy. All patients were anemic with a median hematocrit of 27.5%; nine patients were thrombocytopenic with a median platelet count of 46,0OO/pl. There was no surgical mortality and minimal morbidity. Median hematocrit was 35% at 1 month and 6 months postsplenectomy. Median platelet count in previously thrombocytopenic patients was 261,0OO/p1 at 1 month and 177,OOO/pI at 6 months postsplenectomy. Median survival postsplenectomy in this advanced patient group was 24 months with four patients surviving more than 30 months. Sple-nectomy is a valuable adjunct to the management of end-stage CLL. Cancer S9:1815-1818

cytes. Although patients may not require therapy at presentation, the disease eventually progresses and the development of anemia and thrombocytopenia necessitates treatment. Primary treatment modalities include corticosteroids and cytotoxic drugs as well as radiotherapy. Whereas partial responses generally occur, complete remissions are uncommon and cures are not documented. Therefore, nearly all patients become candidates for salvage therapy .
No optimal salvage therapy has been established for CLL. Since patients in need of salvage therapy often have splenomegaly with some element of hypersplenism, splenectomy is a logical option. However, since marrow failure is often present as well, the value of splenectomy may be limited. Nevertheless, some clinical studies have suggested a value for splenectomy in treating the cytopenias which characterize the late stages of refractory CLL. 1-3 To further define the role of splenectomy in CLL and to assess the impact of splenectomy on survival in refractory CLL, we reviewed our clinical experience with 13 patients who Hospital, Nashville, Tennessee. stitute.

Award.
Accepted for publication December 29, 1986. underwent splenectomy as salvage therapy for advanced CLL.

Patients and Methods
The records of 13 patients with CLL, who were diagnosed by standard criteria and splenectomized at Vanderbilt University Hospital between February 1980 and December 1985, were reviewed. Three patients with CLL and concurrent idiopathic thrombocytopenic purpura or autoimmune hemolytic anemia, who underwent splenectomy for treatment of autoimmune disease during this same time period, were not included in this analysis.
All patients had received chemotherapy before splenectomy. The most common regimen was intermittent chlorambucil and prednisone (nine patients). One patient had received only prednisone, and the others had received combination chemotherapy including an alkylating agent, prednisone and vincristine.
All patients were staged at diagnosis and at the time of splenectomy according to the Rai clas~ification.~ By this classification patients are considered to be in Stage I11 once the hematocrit is <30% or the hemoglobin is < I 1 g. Stage IV disease is defined as a platelet count < 100,000/p1. Anemia and thrombocytopenia which had become refractory to alkylating agents and steroids, and the inability to tolerate further cytotoxic chemotherapy because of cytopenias were the major indications for splenectomy in all patients.
Survival was calculated from diagnosis, from the time of development of Stage I11 disease, and from the time of splenectomy. Actuarial survival was calculated using the method of Kaplan and Meier.'

Patient Characteristics
Information regarding clinical features as well as time from diagnosis to splenectomy and survival after splenectomy are presented in Table 1. The median age at diagnosis was 56 years (range, 49 to 7 1 years). The median interval from diagnosis to splenectomy was 75 months (range, 25 to 188 months). The median interval from diagnosis of Stage 111 disease to splenectomy was 32 months (range, 2 to 135 months).

Splenic Pathologic Findings
All spleens showed extensive involvement by leukemia. Spleen weights ranged from 1000 to 3572 g. Eight spleens weighed between 1000 g to 2000 g; two weighed between 2000 g and 3000 g , and two weighed more than 3000 g .

Operative Complications
No deaths occurred in this series. Three complications occurred: one subphrenic abscess, one wound infection, and one pleural effusion.

Survival
As shown in Figure 1 of survival was fairly typical for CLL in that there was a long period (approximately, 80 months) associated with minimal mortality, followed by a period in which fatal events were frequent. Despite the fact that all patients were in Stage 111 or IV at the time of splenectomy, median actuarial survival after splenectomy was 24 months (Fig.  2). Four patients survived beyond 30 months, and three patients remain alive (14, 24, 32 months, respectively).

Hematologic Response
All 13 patients were anemic (hematocrit, <30%) before splenectomy. At 1 month after splenectomy, all patients showed a favorable response, i. e., hematocrit greater than 30% with no patients requiring transfusions. Eleven patients were alive 6 months after splenectomy; eight of these 11 patients still had hematocrits greater than 30%. The median baseline hematocrit was 27.5%. Median hematocrit at 1 month and 6 months after splenectomy was 35%.
Nine patients were thrombocytopenic, platelets < 100,000/~1 before splenectomy. All nine had platelet counts greater than lOO,OOO/pl at 1 month after splenectomy. Seven of these patients survived at least 6 months after splenectomy, and six of these seven patients had a platelet count greater than 100,OOO/pl at that time.

No. 10
Among patients who were initially thrombocytopenic, the median initial platelet count was 46,OOO/pl. At 1 month postsplenectomy the median platelet count in these patients was 26 1 ,OOO/pl. At 6 months postsplenectomy the median platelet count in the seven survivors who were initially thrombocytopenic was 177,00O/pl.
Although splenectomies were performed to increase the patients' ability to tolerate chemotherapy, because of these hematologic responses, 1 1 of 13 patients received no therapy other than tapering doses of corticosteroids during the 6 months after splenectomy. Because of this high rate of response to splenectomy and the small size of our series, we were unable to determine any prognostic features which predicted either a favorable or unfavorable response to splenectomy.

Discussion
Since therapy of CLL is palliative, all patients with CLL eventually become candidates for further therapy after initial treatment with alkylating agents and prednisone. Various authors have studied whole-body irradiatioq6 splenic irradiation,' thymic irradiation,' and additional chemotherapy' in these patients. One problem with all of these therapies, is that they are myelotoxic, and the patient with advanced refractory CLL may be cytopenic with limited bone marrow reserve.
Although marrow fadure is clearly a contributing factor to the anemia and thrombocytopenia of patients with refractory CLL, splenomegaly is often present. Therefore, it is logical to assume that hypersplenism may be contributing to the cytopenias in these patients. Splenectomy is, therefore, a rational option in the patient with advanced CLL, and previous studies have reported favorable responses in this clinical situation.
In 1975, Adler et al.' reported a series of 50 patients (20 with CLL and 30 with lymphoma) who underwent splenectomy. Despite marrow involvement by lymphoma in most patients, 40 of 48 evaluable patients had a favorable hematologic response. Surgical mortality was only 8%. Median survival after splenectomy was 14.6 months. However, it is difficult to compare that result with the current series since CLL patients and lymphoma patients were grouped together. Additionally, the timing of splenectomy relative to the course of disease in these patients was variable. Nineteen patients underwent splenectomy within 1 year of diagnosis, whereas 16 patients underwent splenectomy 3 to 5 years after diagnosis. Although the patients had advanced disease, the interval from development of advanced disease to splenectomy was not specified.
In 1977, Christensen ez ul.' confirmed the value of splenectomy in patients with CLL. In 22 patients with anemia requiring transfusions, the need for transfusions was reduced in 17. Improvements in platelet counts were also noted. However, this study did not provide any information regarding survival after splenectomy.
Merl et aL3 have confirmed the favorable effect of splenectomy on thrombocytopenia in CLL. However, since median follow-up was only 9 months, the effect of splenectomy on survival cannot be assessed.
In this article we report 13 patients who underwent splenectomy for advanced, refractory CLL. Ten patients had Stage IV disease whereas the others had Stage 111 disease. The median time from the development of Stage 111 disease to splenectomy was 32 months.
As in other series, the hematologic response to splenectomy was good. In all patients, the hematocrit was maintained at or above 30% and the platelet count was above 1 OO,OOO/pl 1 month after splenectomy. Six months after splenectomy, eight of 11 evaluable patients had a hematocrit >30% and only one patient had had a recurrence of thrombocytopenia.
Although splenectomies were performed to increase the patients' ability to tolerate chemotherapy, in the 6 months after splenectomy, 1 1 of 13 patients required no therapy other than tapering doses of prednisone. There was no surgical mortality in the series and morbidity was minimal.
Despite the fact that splenectomies were performed late in the course of disease, the duration of survival after splenectomy was surprisingly good. Survival in CLL is dependent on stage, with survival after development of anemia or thrombocytopenia generally being less than 2 year^.^.'" Our patients had had advanced disease (at least Stage 111) for a median of 32 months at the time of splenectomy and were refractory to chemotherapy (or at least unable to tolerate effective doses of chemotherapy because of cytopenias). Clearly, these patients could be regarded as end-stage patients.
Nevertheless, after splenectomy, the median survival in our series was 24 months with four patients surviving beyond 30 months and three remaining alive (1 4,24,32+ months, respectively). We cannot exclude the possibility that these good results are dependent on the fact that this patient group was somehow favorably selected. Indeed, despite the presence of advanced disease, these patients were considered well enough to undergo splenectomy. However, we believe that the survival which we observed after splenectomy represents a beneficial effect of the procedure. Additionally, correction of cytopenias may have limited the morbidity associated with CLL, and may have improved tolerance of chemotherapy in the terminal phases of the disease.
Based on our findings, we would recommend that splenectomy should be considered in any patient with Stage 111 or 1V CLL, especially when progressive cytopenias compromise the administration of cytotoxic agents. Whereas splenectomy might be more effective if employed