Thrombocytosis and Thrombocythemia: The Laboratory and Clinical Significance of an Elevated Platelet Count

Thrombocytosis and Thrombocythemia: The Laboratory and Clinical Significance of an Elevated Platelet Count WILLIAM M. DAVIS, M.D., AND ALICIA O. MENDEZ ROSS, M.D. Departments of Medicine and Pathology, Hematology Service and Hematology Laboratory, Springfield Hospital Medical Center, Springfield, Massachusetts 01107 ABSTRACT Davis, William M., and Ross, Alicia O. Mendez: Thrombocytosis and thrombocythemia: The laboratory and clinical significance of an elevated platelet count. Am. J. Clin. Pathol. 59: 243-247, 1973. One hundred consecutive patients with elevated platelet counts (more than 500,000 per cu. mm.) seen in a 16-month period have been studied. Thirty-six were found to have underlying malignant disease. Fifteen of the patients had recently undergone splenectomy, and four patients had recently had other major surgery. Five of the patients had polycythemia vera, and one patient had chronic granulocytic leukemia. Iron deficiency anemia was associated with thrombocytosis in eight patients. There were 24 patients with miscellaneous benign conditions and seven patients with collagen disorders, usually rheumatoid arthritis. The thrombocytosis of malignant disease was frequently unexpected and was an important clue to the presence of a disseminated malignancy. AN ELEVATED PLATELET COUNT has been re- underlying diseases. On the other hand, ported in a number of clinical conditions. 2 - thrombocythemia refers to a marked and 4-7, D, ii, i2, i4, IB This elevation undoubtedly persistent elevation of the platelet count reflects an increase in the rate of thrombo- as part of a myeloproliferative disorder, cyte production and release rather than a often with systemic manifestations, prolongation of the platelet life span. The Thrombocytosis associated with a numspecific stimulus for such production in- ber of diseases is well known to most phycrease or release is still unknown. How- sicians. Thrombocytosis has been docuever, several investigators have suggested a mented in diseases such as chronic inflamhumoral mechanism. l > 8-10 - 1S The concept matory bowel disease, 7 rheumatoid arthriof a feedback mechanism 1 has also been tis, acute infections, and trauma, and has suggested. also been noted following major surgical Ginsburg and Aster 5 have defined throm- procedures, 4 particularly splenectomy. Fibocytosis as a moderate, short-lived, and nally, thrombocytosis may be the clue of symptomless elevation of platelet count an underlying malignancy in some patients that occurs in association with certain in whom the diagnosis is not yet suspected or established. 5 A re-evaluation of the signia R n e u C stip d t MarchTo, IW^SfS? pus lion March 29, 1972. "ificance of an elevated platelet count in a large general hospital seemed desirable. 243

244 DAVIS AND ROSS A..J.C.P. Vol. 59 Methods and Results One hundred consecutive patients with elevated platelet counts (more than 500,000 per cu. mm.) seen in a 475-bed general hospital during a 16-month period from May 1968 through August 1969 were studied. During this period of time the Hematology Laboratory performed 3,642 platelet counts on blood from 907 patients, from whom these 100 consecutive patients were culled. Inpatients as well as outpatients were included in the study. During the 16- month period there were 17,439 patient admissions. Platelets were counted by phase contrast microscopy by the method of Brecher and Cronkite. 3 These authors have reported the mean platelet count as 257,000 per cu. mm. The distribution of these counts was skewed to the right and statistical estimation of the 95% limits of variation of such single counts was 140,000 to 440,000 per cu. mm. Sloan 16 reports a similar mean and range. We have accepted that the normal platelet count ranges from 150,000 to 450,000 per cu. mm. For the purpose of this study, any patient with a platelet count of 500,000 per cu. mm. or more was included. The observed platelet counts ranged from 500,000 to 1,500,000 per cu. mm. The initial clue to the presence of thrombocytosis was often the review of the peripheral blood smear. Thus, the platelet count was frequently carried out because the technician had reported the number of platelets to be "increased." In this study there were 60 women and 40 men. There were 15 patients from 0 to 9 years of age and 33 patients in the fourth and fifth decades of life. It is probable that this pattern represents the hospital population age distribution. Table 1 shows the major causes of thrombocytosis in this study. The largest group of patients with thrombocytosis were the 36 patients with malignancies. The breakdown as to primary tumors was as follows: carcinoma of the ovary, seven; carcinoma of the breast, five; carcinoma of the lung, four; carcinoma of the pancreas, three; carcinoma of the colon, four; metastatic adenocarcinoma, primary site undetermined, two; miscellaneous carcinoma, two; malignant melanoma, two; malignant lymphoma (including Hodgkin's disease), seven. In none of our patients did we find early small localized tumors; in almost all cases the malignancy was disseminated. However, the presence of malignant disease was often not strongly suspected. There were six patients with myeloproliferative disease. Five patients had polycythemia vera and one patient had chronic granulocytic leukemia. During this period of study, no patients with myelofibrosis or essential thrombocythemia were observed. Among those patients with thrombocytosis occurring after surgical procedures, the most common operation was recent splenectomy in 15 patients as follows: in four adult patients with chronic idiopathic thrombocytopenia purpura (early thrombocytosis represented an immediate therapeutic response); in five patients because of trauma; and in six patients for hemolytic anemias of several types. There were four patients who had recently had major surgery as follows: gastrectomy for massive gastrointestinal bleeding, bowel resection for chronic inflammatory bowel disease, craniotomy for subarachnoid hemorrhage, and cholecystectomy for chronic cholecystitis. There were eight cases of well-documented iron deficiency anemia; five of these patients were in the pediatric age group. None of the three adult patients had underlying malignancy. There were seven patients with collagen disorders, usually rheumatoid arthritis. There were 24 patients with miscellaneous disorders: eight had acute infection, five had cardiac disease, and five had either chronic pancreatitis or cirrhosis of the liver, or both.

February 1973 SIGNIFICANCE OF THROMBOCYTOSIS 245 One patient admitted with anemia, leukopenia, and severe thrombocytopenia (platelet count less than 10,000 per cu. mm.) was found to have Addisonian pernicious anemia. During the immediate response to vitamin B 12 therapy the platelet count rose to 600,000 per cu. mm. A 41- year-old woman was admitted with a left hemiplegia and found to have congenital spherocytic hemolytic anemia. She had not had a splenectomy. As part of the active hemolytic process, presumably, the platelet count was elevated to 1,018,000 per cu. mm. The patient refused the recommended splenectomy. The records of all these patients were carefully reviewed for possible associated thrombotic or hemorrhagic complications. The most striking elevations in platelet counts occurred in the patients immediately post-splenectomy. There were no complications of this type in any of these patients. No therapy such as heparinization was used. In only one of the 36 patients with disseminated malignant disease was a thrombotic episode observed. This patient, with carcinoma of the ovary, showed bilateral pulmonary emboli and multiple areas of pulmonary infarction at postmortem examination. The paradoxical risk of both hemorrhage and thrombosis in patients with polycythemia and the myeloproliferative disorders is well established. One of our patients with untreated polycythemia presented with a left hemiplegia clue to cerebral thrombosis. In addition, one patient with polycythemia vera at the time of surgery demonstrated significant postoperative bleeding, primarily from the upper gastrointestinal tract. Discussion Since Riess 11 first observed the association of malignancy and thrombocytosis in 1872, several investigators have reported similar findings. However, these early re- Table 1. Causes of Thrombocytosis Disease or Condition No. of Patients Malignancy 36 Polycythemia vera 5 Chronic granulocytic leukemia 1 Surgery Splenectomy 15 Other 4 Iron deficiency anemia S Klieumatoid arthritis or other collagen disorders 7 Miscellaneous 24 TOTAL 100 ports appeared principally in the European literature. 2 ' u More reports, however, are now appearing in the U. S. medical journals. 6 ' 14 ' 15 One of the recent reports was the study of Levin and Conley, 6 who in 1964 surveyed 82 patients with elevated platelet counts occurring in a large group of 14,000 patients. An elevated platelet count in their study was considered to be 400,000 per cu. mm. or more. Of these 82 patients, 31 (38%) had underlying neoplasms. These authors subsequently studied another group of 240 consecutive patients with known inoperable malignancies, prior to any radiation or chemotherapy. The incidence of thrombocytosis in this series was 45%, excluding those with carcinoma of the lung; patients with carcinoma of the lung had a 38% incidence of thrombocytosis. In a more recent study, Silvis and associates 14 reviewed 190 patients with proven carcinoma of the lung and noted that 60% of these patients had significant thrombocytosis. Thrombocythemia associated with myeloproliferative disorders is well recognized, particularly polycythemia vera and chronic granulocytic leukemia in the early phases. As chronic granulocytic leukemia evolves, and partly as a result of treatment, the

246 DAVIS AND ROSS A.J.C.P. Vol. 59 platelet count eventually falls to the normal range and later will be significantly depressed. Myelofibrosis at some stages will often be accompanied by a significant thrombocytosis. The fourth member of the myeloproliferative group, essential thrombocythemia, is rare. In this group of patients with myeloproliferative disorders there will usually be significant platelet morphologic abnormalities, particularly variation in size and the presence of giant forms. Recent surgery as a cause of increased platelet counts is also well established. The recent study of Breslow and co-workers* showed that this thrombocytosis could occur after a variety of major surgical procedures but was most frequent after splenectomy. After major procedures other than splenectomy there was often an initial decrease in platelet count followed by elevation to supranormal levels. The thrombocytosis was associated with an increase in circulating megakaryocytes and megakaryocytic fragments caused by an unknown mechanism. The increase after splenectomy was often seen immediately postoperatively, but sometimes it was delayed significantly for 3 to 10 days. The platelet counts after splenectomy usually peaked at about 3 to 4 weeks, but tended to persist for many months and years after the operation. Various benign conditions have been associated with elevated platelet counts. This was discussed by Silvis and co-workers 15 in their review of 322 patients with thrombocytosis. Of these, 116 cases (36%) were associated with "benign" conditions such as iron deficiency, recent hemorrhage, chronic inflammatory bowel disease, cirrhosis, alcoholism, rheumatoid arthritis, pernicious anemia (as part of the response to initial therapy), and tuberculosis. Thrombocytosis seen in patients responding to vitamin B ]2 therapy for Addisonian pernicious anemia has been documented by Ogston and Dawson." The increase in platelets began with and corresponded to the start of the reticulocyte response. The peak platelet count was usually reached about 9 to 12 days later. The association with iron deficiency has been observed by a number of investigators. With effective treatment of iron deficiency, the platelet count will return to normal. An additional observation of note is that platelets of iron deficiency are often small and thus "microcytic." Acute bleeding (major hemorrhage) is often followed by thrombocytosis. The mechanism of thrombocytosis in most patients with benign conditions is unknown, and the frequency of its occurrence is not high. Our group of patients with benign disorders was similar to those reported in the literature. The possibility that the elevation of the platelet count will be associated with either bleeding or thrombotic manifestations must be considered. Despite the relatively abrupt elevations in platelet counts occurring shortly after splenectomy, none of the patients showed evidence of thrombotic phenomena. Two instances of cerebral thrombosis occurred in one patient with congenital spherocytic hemolytic anemia (who had not had a splenectomy) and another patient with early untreated polycythemia vera. The medical literature is replete with documentation of the bleeding and thrombotic complications of the myeloproliferative disorders, particularly polycythemia vera. The bleeding observed appears to be related to abnormal platelet function in part, although, in polycythemia, blood viscosity and vascular factors are also important. The thrombotic complications are probably related to blood viscosity and vascular changes of aging rather than to the increased platelets. It would not appear that an elevation in platelet count per se requires therapy. However, in the myeloproliferative disorders the disease itself frequently needs therapy with alkylating

February 1913 SIGNIFICANCE OF THROMBOCYTOSIS 247 agents or radioactive phosphorus, which will almost always lower the platelet count significantly. In many clinical situations thrombocytosis or thrombocythemia is an expected finding. The post-splenectomy patients and those with the myeloproliferative disorders are in this category. In this series there were 19 post-surgery and post-splenectomy patients and six patients with myeloproliferative disorders. In these 25 patients the elevated platelet counts were expected. In the remaining 75 patients, including 36 patients with malignancies, the thrombocytosis could be said to have been unexpected. Thus, an unexplained thrombocytosis would appear to be associated with a 48% incidence or approximate 50/50 chance, of malignancy. This finding of a platelet "increase" or a platelet count in excess of 450,000 to 500,000 per cu. mm. has real clinical significance for the patient and his physician. References 1. Abilgaard CF, Simone VJ: Thrombopoiesis. Semin Hematol 4:424-452, 1967 2. Bagdasarov AA, Rodina RI, Gefen EG: Blood platelets in cancer and peptic ulcer. Klin Med 28:39-45, 1950 3. Brecher G, Cronkite EP: Morphology and enumeration of human blood platelets. J Appl Physiol 3:365-377, 1950 4. Breslow A, Kaufman RM, Lawsky AR: I'he effect of surgery on the concentration of circulating megakaryocytes and platelets. Blood 32: 393-401, 1968 5. Ginsburg AD, Aster RH: Platelet function: Its clinical significance. DM, Sept, 1970, pp 36-39 6. Levin J, Conley CL: Thrombocytosis associated with malignant disease. Arch Intern Med 114: 497-500, 1964 7. Morowitz AD, Allen LW, Kirsner JB: Thrombocytosis in chronic inflammatory bowel disease. Ann Intern Med 68:1013-1021, 1968 8. Odell TT Jr, McDonald TP, Detwiller TC: Stimulation of platelet production by serum of platelet depleted rats. Proc Soc Exp Biol Med 108:428-431, 1961 9. Ogston D, Dawson AA: Thrombocytosis following thrombocytopenia in man. Postgrad Med J 45:754-756, 1969 10. Pierre RV, Linman JW: Studies on thrombopoiesis. IV. Thrombocytosis-promoting activity of normal urine. Proc Soc Exp Biol Med 113: 398-400, 1963 11. Pranis LIU: Thrombocytopoiesis in cancer and pre-cancer of the stomach according to data on the peripheral blood and the bone marrow. Vopr Klin Lech Zlok Novoobraz 4:137-150, 1956 12. Riess L: Zur pathologischen anatomic des blutes. Arch Anat Physiol Wissensch Med, 1872, pp 237-249 13. Schulman I, Pierce M, Lukens A, et al: Studies on thrombopoiesis. I. A factor in normal human plasma required for platelet production; chronic thrombocytopenia due to its deficiency. Blood 16:943-957, 1960 14. Silvis SE, Turkbas N, Doscherliolmen A: Thrombocytosis in patients with lung cancer. JAMA 211:1852-1953, 1970 15. Silvis SE, Turkbas N, Swaim WR, et al: Causes of thrombocytosis: A clinical evaluation of 322 patients. Minn Med 52:1603-1607, 1969 16. Sloan AW: The normal platelet count. J Clin Pathol 4:37-46, 1951