Myeloproliferative Disorders

SEYMOUR 1. SCHWARTZ, M.D.*
Forty-three operative procedures were performed on a population of 250 patients with myeloproliferative disorders, including polycythemia vera, myeloid metaplasia (MM) and chronic myelogenous leukemia (CML). The overall operative mortality was approximately 7% and the incidence of excessive bleeding which could be related to coagulopathy was 5%. Twenty-one patients with MM or CML underwent splenectomy for palliation of symptoms related to the enlarged spleen or hematologic prob- lems. Eighty-four percent of the latter group were improved. Adverse hematologic effects which could be attributed to splenec- tomy in these patients were confined to two patients who de- veloped marked thrombocytosis. Among the 23 patients with MM, 9 had portal hypertension. Three underwent portacaval shunt and one a splenorenal shunt for bleeding varices. One of the patients died of hepatic necrosis. Estimated hepatic blood flow determinations (EHBF) in 4 patients with portal hypertension demonstrated a marked absolute increase and an increase in the ratio of EHBF/Cardiac Index. Absence of any evidence of in- trahepatic or extrahepatic obstruction in these patients and the demonstration that splenectomy relieved portal hypertension de- fined at surgery in 4 patients, suggests that augmented adhepatic flow contributes to portal hypertension in some cases. The review leads to the conclusions that: 1) Operative procedures in prepared patients with myeloproliferative disorders are not associated with prohibitive mortality and morbidity rates. 2) Splenectomy is indi- cated for patients with increasing transfusion requirements and symptomatic splenomegaly or hypersplenism and should be per- formed early in the course of disease. 3) When associated portal hypertension and bleeding varices are present, hemodynamic studies should be carried out to define if splenectomy alone, or a portal systemic decompressive procedure is indicated.
THE MYELOPROLIFERATIVE DISORDERS include a spec- s trum of diseases, each of which may be due to a
different disturbance in proliferation and/or maturation of the hematopoietic stem cell of marrow. Polycythemia
From the University of Rochester School of Medicine and Dentistry, and Strong Memorial Hospital,
Rochester, New York
Vera (PV), Acute and Chronic Myelogenous Leukemia (AML and CML), Myeloid Metaplasia (Myelofibrosis, Myelosclerosis) (MM) and Essential Thrombocythemia all have been included within the general category of myeloproliferative disorders. Surgical consideration of patients with myeloproliferative disorders relates to: 1) a general appreciation of the operative risk for any surgical procedure in these patients; 2) the application of splenec- tomy as a palliative procedure in patients with MM and CML; and 3) the management of portal hypertension and variceal bleeding in patients with MM.
Material and Methods The records of 250 patients with myeloproliferative
disorders seen at the Strong Memorial Hospital between 1955 and 1974 were reviewed. Included were 90 patients with PV, 64 patients with MM and 96 patients with CML. Some of these patients have progressed from PV to MM. Each patient was assigned to the more advanced categ- ory, unless an operative procedure was performed, in which case he or she was considered in that category during which the operation was carried out. All patients had the diagnosis established by blood counts, peripheral blood smears, bone marrow aspirates or biopsies. In pa- tients subjected to splenectomy and/or portal decompres- sive procedures, histologic evaluation of the spleen and/or liver contributed to the diagnosis. A
Operative mortality was considered as that occurring during the hospitalization for the procedure or within 30 days. Excessive bleeding related to an hemostatic defect was defined as that associated with a 5% reduction in
464
Presented at the Annual Meeting of the American Surgical Associa- tion, Quebec City, Quebec, May 7-9, 1975.
*The University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue, Rochester, New York 14642.
MYELOPROLIFERATIVE DISORDERS
hematocrit, requirement for a number of transfusions significantly greater than that usually associated with the procedure, reexploration at which a single bleeding site could not be defined, or a combination of the above.
Portal pressures were measured during the operative procedure in either the portal vein itself, or in an omental vein with the portal vein as a reference level, using a spinal manometer filled with saline. Central venous pres- sures (CVP) were determined simultaneously. Hemodynamic studies were performed on awake pa-
tients who were not premedicated. Hepatic blood flow (EHBF) during the postabsorptive state was carried out according to the method of Leevy et al.12 employing indocyanine green dye. Pressure measurements in these patients were made using Statham P-23 AA strain gauges and a Sanborn direct writing recorder. Cardiac output determinations were made from the standard indocyanine green indicator dilution curves. Body surface area esti- ,mates were derived from a Du Bois normogram.
Results I. Operative procedures in all categories
Forty-five operative procedures were performed in 41 ,patients. Twelve procedures were carried out in patients with PV, 29 were performed in patients with MM, and four procedures were performed in patients with CML (Table I). There were 4 deaths related to the operative procedure, resulting in an overall mortality rate of 9o. None of the patients with PV died; one of the patients with CML died, while in the group with MM, the mortal- ity rate was 10%o (Fig. 1).
In the MM group, a patient died following splenectomy with ensuing subphrenic abscess and sepsis. Another pa-
tient died due to hepatic necrosis following an end-to-side portacaval anastomosis. The third death in the MM group
occurred in a 48-year-old woman who had a 4,950 gram spleen removed because of symptomatic splenomegaly on 5-18-72 without intraoperative or postoperative trans- fusional requirement. On 2-28-73 she underwent trans-
OPERATIVE PROCEDURES IN PATIENTS WITH MYELOPROLIFERATIVE DISORDERS
TOTAL CASES
OPERATIVE CASES
PERCENT MORTALITY FIG. 1. Mortality Associated with Operative Procedures in patients with Myeloproliferative Disorders.
thoracic subclavian endarterectomy. Her preoperative hematocrit was 35, WBC 50,000 and platelets 160,000/ mm.3 Postoperatively she developed an hemothorax which required three reexplorations. She subsequently developed aspiration pneumonia, respiratory failure and died 6 weeks after the first thoracotomy. One patient who underwent splenectomy for hypersplenism associated with CML died following discharge but within one month of surgery. At autopsy there were extensive thromboses of the renal veins and mesenteric veins.
In the group of patients with PV, there was no exces- sive bleeding, and only one significant postoperative in- fection; a subphrenic abscess following cholecystectomy for acute cholecystitis. In one of the patients, the preoperative platelet count was 843,000/mm3 while in all others this was in the normal range. One patient was
TABLE 1. Operative Procedures in Patients With Myeloproliferative Disorders
PV MM CML
Arterial Reconstructions 3 Splenectomy 22 (1*, 1+) Splenectomy 3 (1+) Amputations 2 (With Splenorenal shunt 1) Vagotomy & Antrectomy 'Lumbar Sympathectomy Portacaval Shunt 3 (1+) Aortic Valve Replacement Subclavian Endarterectomy 1 (1+) Pericardial Poudrage Cholecystectomy Nephrectomy Vagotomy & Pyloroplasty Mastectomy Hernia Repair Drainage Subphrenic Abscess Cholecystectomy
Total: 12 Total: 29 Total: 4 Deaths: 0 Deaths: 3 Deaths:
*Excessive bleeding + Death
Ann. Siurg. - October 1975SCHWARTZ466 TABLE 2. Indications for SplenectomylEffects
Improvement Improvement Primary In Survivors Secondary In Survivors
Symptomatic Splenomegaly/ Hypermetabolism 8 (2+) 6/6 Anemia/
Hemolysis 10 7/10 6 5/6 Thrombocyto-
penia 4 3/4 8 7/8 Leukopenia 2 2/2
Total: 24 19/22 (86%) 12/14 (86%)
+ Death
operated on with a hematocrit of 63, while in the remain- der the hematocrit was less than 52%. Ten of the 12 patients had been treated with busulfan, 32P, multiple phlebotomies or a combination of these modalities. In the 6 patients requiring an operation for peripheral arterial disease, neither hematocrit nor platelet count were signif- icantly increased. Among the 27 patients with MM undergoing 29 proce-
dures, excessive bleeding occurred in 3 cases; in the subclavian endarterectomy described above and in two of the 21 splenectomies. In none of these three could the bleeding be attributed to a specific bleeding site, and none of these patients had platelet counts less than 70,000/mm.3 In both splenectomy cases, reexploration and evacuation of the clot from the left subphrenic space resulted in cessation of bleeding. There was no excessive bleeding in the two patients with platelet counts greater than 350,000/mm.3 At the operative procedure in patients with MM, no patient had a hematocrit greater than 52. Two of the three deaths in the group of patients with MM were related to sepsis. The three splenectomies in the CML group were per-
formed in patients with mild anemia; two had throm- bocytopenia, and one mild thrombocytosis. There were no immediate postoperative complications, but one pa- tient had a slightly elevated platelet count and died within a month of surgery from extensive visceral venous thromboses and a blastic crisis.
II. Splenectomy as a Palliative Procedure
Twenty-four splenectomies were performed as pallia- tive procedures in patients with MM or CML. The indica- tions are listed in Table 2. In 8 patients the major indica- tion was symptomatic splenomegaly in which the spleen was of such magnitude as to be uncomfortable or to interfere with alimentation, or in which multiple episodes of painful splenic infarcts occurred. In some patients there was associated hypermetabolism. In 10 patients, a significantly increased transfusion requirement or hemolysis was the major indication, while in 4 patients
thrombocytopenia with clinical manifestations consti- tuted the major indication. In two patients leukopenia interfering with the defense mechanism toward an infec- tion, prompted splenectomy. All patients had received medical therapy, usually from months to years prior to the operation. The regimen varied and consisted of phlebotomies, prednisone, :'2P, busulfan, cyclophos- phamide or a combination of these. The mean weight of the spleens which were removed
for symptomatic splenomegaly was about 3,150 gm, range of 2,250-7,650 gm. There were two patients with an accessory spleen. Five of the patients had platelet counts of 70,000/mm:3 or less. The left subphrenic space was drained in 7 patients. Reexploration for continued bleed- ing was required in 2 patients, one of whom had throm- bocytopenia, and one was a patient in whom the space was not drained. One of the patients developed a sub- phrenic abscess and died of sepsis a month after the second procedure. The other death occurred in the pa- tient with CML described above. The 6 remaining pa- tients had symptomatic improvement, and in no instance was there an adverse hematologic effect, i.e., more pro- found anemia, thrombocytopenia or leukopenia. Two of the patients have died, years after the operative proce- dure for CML.
In 16 patients, the major indication was hypersplenism, with anemia, thrombocytopenia or leukopenia the factor prompting the operative procedure (Table 2). The mean weight of the spleens in these patients was 2,250 gm (range 675 gm-3,150 gm). There were three patients in whom small, accessory spleens were also removed. The subphrenic space was drained in 14 patients. The hematologic defect which constituted the prime
indication for surgery was improved in 12 patients for an overall therapeutic effect of 75% (Fig. 2). Seven of the 10 patients with an increase in transfusion requirements had this significantly reduced for months to years following splenectomy. Five of the 6 patients in whom anemia was a contributory factor, had some increase in the hemato- crit following splenectomy. Three of the 4 patients with symptomatic thrombocytopenia as the major factor, im- proved following splenectomy. Platelets were not ad-
INDICATION 6 CASES
FIG. 2. Results with Spienectomy in Patients with Myeloproliferative Disorders.
\o MYELoPROLIFERATIVE DISORDERS
ministered preopei-atively in these patients but 10 to 20 packs were given intraoperatively to 2 of the patients following splenectomy. In one of these. there was sig- nificant postoperative bleeding requiring evacuation of a hematoma. but the platelet count increased significantly from 1 ,000/mm' to 60.000/mmT;3 the week after surgery and was maintained above 70.000/mm: for 5 years. The pa- tient in whom no improvement occurred, has a platelet count which has persisted at 18,000/mm' and has had no bleeding since surgery for the past 4 years. In the 8 patients with thrombocytopenia as a contributory factor, an increase in the count of at least 20,000/mm' has been maintained in 7 patients since splenectomy. The two patients subjected to splenectomy for uncon-
trollable pulmonary infection related to white blood counts of 800/mm' and 1200/mm:', had spleen weights of 675 gm and 1,350 gm respectively. There were no acces- sory spleens. In both instances, the infection came under control postoperatively with continuance of preoperative antibiotic therapy which had previously been ineffective. In one patient. the initial increase in white blood count was minimal, i.e., 800/mm3 to 1600/mm3, and the white blood count has been sustained at about 3,000/mm3 with no subsequent infections for the past month. In the second patient, the white blood count increased to 6,000/ mm3 and has been maintained for 6 months at that level without evidence of further infection. There were two significant postoperative infections;
both subphrenic abscesses. One of these patients died of uncontrollable sepsis. The other developed gram nega- tive sepsis due to Se1rra(ati inar(cescens following drainage of an infected subphrenic hematoma. This responded to gentamycin and the patient did well for 41/2 years. There was no instance of worsening of the hematologic profile subsequent to splenectomy in ant patients who under- went splenectomy for palliation. The WBC increased in the immediate postoperative period in 84% of the pa- tients. In I I there was marked leukocytosis preopera- tively. and this persisted postoperatively. In two patients there were platelet counts of 480,000/mm3 and 380,000/ mm3 preoperatively. In the former, the count increased temporarily to 843.000/mm' with no adverse effects. The one death which may have been related to throm- bocytosis occurred one month after surgery in a patient with CML. His maximum platelet count was 420,000/ mm3, but at autopsy there were thromboses and multiple mesenteric and renal veins. Three of the patients with MM and one of the CML
patients have died of their disease 3 to 7 years following splenectomy. In two of these, infection, in one a case of Crvptococcus neofo'rnanls, contributed to the demise.
III. Myeloid Metaplasia and Portal Hypertension Four patients with MM underwent decompressive pro-
467 TABLE 3. Portal Decompressii'e Procedures In Patients With Myeloid
Metaplasia
Splenorenal Shunt F, 60 + Died 3 years later-pneumonia
Mild recurrent encephalopathy End-to-Side
Portacaval shunt F. 41 - Died 36 hours later Massive hepatic necrosis
it M, 77 + Died 6 years later-M.I. Mild Recurrent encephalopathy
F. 74 - Died 3 years later-? Etiology Mild recurrent encephalopathy
cedures for portal hypertension, esophagogastric varices and a history of recurrent upper gastrointestinal bleeding (Table 3). One 60-year-old woman who had been treated for pernicious anemia at age 38 and who had been diag- nosed at age 50 as having MM, underwent splenectomy and splenorenal anastomosis in 1966. In the 10 antecedent years she had had multiple phlebotomies, splenic irradia- tion and several courses of busulfan. At surgery the liver biopsy revealed portal cirrhosis in addition to minor foci of extramedullary hematopoiesis. Since splenectomy alone affected only a minor reduction in portal pressure, a splenorenal shunt was carried out and the blood pres- sure reduced to near caval pressure. Her three year post- operative course was complicated by several episodes of encephalopathy which responded to protein restriction and neomycin. She died of pneumonia, and at autopsy there was no evidence of varices.
In the group subjected to end-to-side portacaval shunts, one patient developed hepatorenal failure due to massive hepatic necrosis, and died on the second post- operative day. In addition to the acute necrosis, the liver revealed foci of extramedullary hematopoiesis but no cirrhosis. The second patient, a 77-year-old man, had no recurrent bleeding post shunt, but did experience mild recurrent encephalopathy and died from cardiac prob- lems 6 years subsequent to his shunting procedure. At the operative procedure, his liver demonstrated both ex- tramedullary hematopoiesis and cirrhosis. The remaining patient who received an end-to-side portacaval shunt for bleeding varices was the subject of a previous case re- port.16 She survived two years and her course was compli- cated by several episodes of mild recurrent en- cephalopathy. This 74-year-old woman had portal hyper- tension, bleeding varices and no evidence of cirrhosis or regenerative nodules. Small foci of extramedullary hematopoiesis were present in the liver. The presence of portal hypertension unassociated with any hepatic lesion which could be indicted as a cause of outflow obstruc- tion, led to a consideration of increased flow through the splenoportal system as a factor in the production of portal
Vol.1 8 No() 4
TABLE 4. Correlation of Intraoperative Pressures (MM Patients With Suspected Portal Hypertension)
Portal Pressure (mm saline) Before After Spleen
Patient Ascites Varices Splenectomy Splenectomy CVP Cirrhosis Weight (gm.)
I + - 110 - - 1840 2 - + 370 140 80 - 3850 3 + - 170 4 - + 270 110 90 - 3630 5 + + 360 130 110 + 2200 6 - + 480 120 80 - 4950 7 + + 390 320 70 + 2250 8 + + 320 290 90 + 3150
hypertension in the patient with MM and associated two. Similarly, the EHBF was significantly increased in marked splenomegaly. all patients and more than twice normal in two. In only
In 8 patients with MM in whom portal hypertension one of these patients did liver biopsy reveal cirrhosis. was anticipated due to the presence of varices or ascites, portal pressures were determined at the operative proce- DiSCUSSlOi dure, before and after splenectomy (Table 4). Two pa- The presently reported experience with major surgical tients revealed normal pressures in the portal system, procedures in patients with a variety of myeloprolifera- while in 6 an elevation was demonstrated. In 4 of the 6 tive disorders is encouraging in regard to mortality and patients, splenectomy effected a marked reduction in morbidity rates as compared to previous publications. portal pressure to levels within the range of normal but Hemorrhage has been the most commonly reported com- still slightly elevated in reference to CVP. All 4 of these plication, occurring in 75% of the complicated cases, and patients showed only small foci of extramedullary leading to 69% of the death.1521 Venous or arterial hematopoiesis in the liver, and there was no evidence of thrombosis and infection rank second and third on the cirrhosis. In two of these patients, the varices gradually complication list, both in patients with PV and MM. disappeared months after splenectomy. Two patients The hemostatic abnormalities in these patients are with significant cirrhosis experienced only a slight reduc- complex, multifaceted, and not precisely defined. The tion in portal pressure associated with splenectomy. One increased viscosity associated with a high hematocrit is of these was subjected to a concomitant splenorenal thought to lead to tissue hypoxia and consequent inter- shunt. ference with vascular integrity. Quantitative platelet ab- The findings of hemodynamic studies in 4 patients with normalities in the form of thrombocytopenia may con-
MM and suspected portal hypertension are shown in tribute to a compromise of hemostasis, while the less Table 5. The hepatic venous wedge pressure was greater common thrombocytosis may be a factor in thrombosis, than normal in all patients and more than twice normal in bleeding, or a combination of the two. A variety of qual-
TABLE 5. Hemodynamic Findings in Patients With Suspected Portal Hypertension*
Pressures (mm Hg)
Measurement Normal Case 1 Case 2 Case 3 Case 4
Rt. Atrium 0-5 4 2 5 4 Inferior Vena Cava 2 - 7 8 5 13 12 Free Hepatic Vein 4- 10 10 9 14 15 Hepatic Venous Wedge 6- 13 26 15 21 30
Cardiac Output (L/min) 4.97 5.83 8.28 5.92 Cardiac Index (L/min/m2) 2.7 - 5.0 3.23 4.10 3.97 3.92
Estimated Hepatic Blood Flows (ml/min/m2) 480 (±+150) 1485 853 1100 1640
EHBF/C I 0.20 0.46 0.21 0.28 0.42
Varices + + + + Ascites - + - + Cirrhosis - - +
MYELOPROLIFERATIVE DISORDERS
itative platelet abnormalities has been reported.2'1421 These include: viscous metamorphosis,…