Prevalence of HTLV-I-associated T-cell lymphoma

In order to assess the prevalence rate of HTLV-1-associated T-cell lymphomas and human retrovirus infection in general, approximately 21,000 individuals representing various patient populations, retroviral risk groups, and blood donors were examined for HTLV-I, HTLV-II, HIV-1, or HIV-2 infection using serologic and PCR assays. The prevalence rates among volunteer blood donors were 0.02% and 0% for HTLV and HIV, respectively. Significantly increased HTLV prevalence rates were observed among paid blood donors, African American health care clinic patients, Amerindians, recipients of HTLV-positive cellular blood products, intravenous drug users, sexual contacts and family members of HTLV-positive people, and patients with primary thrombocytosis and other-than-low-grade non-Hodgkin's lymphoma (NHL). Among some of these groups there were significant differences in the prevalence of HTLV-I versus HTLV-II. The eight HTLV-positive NHL patients all had mature, high-grade, CD4+ T-cell lymphomas with clonally integrated HTLV-I, for a prevalence of 4% among other-than-low-grade NHL patients. Seven of the eight died from their disease within 2 years despite treatment. Interestingly, two groups at risk for HTLV infection, namely needle stick victims and recipients of HTLV-infected and/or pooled plasma products, showed no evidence for infection. Significantly increased HIV-1 prevalence was observed among paid blood donors, African Americans, homosexuals, female prostitutes, hemophiliacs, and other-than-low-grade NHL patients. Only one patient was infected with HIV-2. Of the nine HIV-positive, other-than-low-grade NHL patients, seven HIV-1 positives had B-cell lymphomas, one HIV-1 positive had an HTLV-I-positive CD4+ T-cell lymphoma, and one infected with HIV-2 had a CD4+ T-cell lymphoma that was HTLV negative. The data indicate that HTLV-I lymphoma, while uncommon, is not necessarily rare among other-than-low-grade NHL cases in the United States and, given its poor prognosis, should probably be studied separately in clinical trials.

Coagulation abnormalities in patients with Gaucher's disease: effect of therapy

Various coagulation defects have been associated with Gaucher's disease, including factor IX deficiency and acquired von Willebrand's disease (VWD). We performed repeated coagulation assays in 9 patients with Gaucher's disease over a period of 2 years. The prothrombin time (PT) and fibrinogen levels were normal in 8 of 9 patients, while the partial thromboplastin time (PTT) was abnormal in 5 of 9; all mixing PTT tests showed correction. Factor IX was normal repeatedly in the 7 of 7 patients tested. In contrast, factor XI was decreased in 3 of 9 patients assayed. Anticardiolipin (ACL) IgM was normal in all patients. ACL IgG was highly variable; levels were abnormal at least once in 6 of 8 patients, but were also normal at least once in 7 of 8 patients. Factor VIII was also quite variable: levels were decreased at least once in 4 of 9 patients, and normal at least once in 8 of 9 patients. Von Willebrand factor antigen (VWF Ag) studies were normal in 7 of 8 patients, but VWF activity was decreased at least once in 4 of 8 patients. In some patients, these problems could be overcome by specimen dilution. In ony 1 patient was VWF Ag decreased; this patient had a factor VIIIC level of 13% , and VWF activity of 18.7%. Coagulation assays performed before and after alglucerase administration failed to demonstrate any significant improvement in these assays, and neither was there a consistent improvement over the duration of therapy. We suggest that previously reported decreases in factor IX and VWF may be secondary to the interfering presence of increased cerebroside levels. Caution must be used in the interpretation of clotting assays in the patient with Gaucher's disease.

Impaired interferon alpha response in hairy cell leukemia is corrected by therapy with 2-chloro-2'-deoxyadenosine: implications for susceptibility to opportunistic infections

Patients with hairy cell leukemia (HCL) are susceptible to opportunistic intracellular infections, suggesting defects in cellular immunity. Prior studies have indicated an association between failure of IFN-alpha generation by peripheral blood mononuclear cells (MNC) and susceptibility to such infections. We here present results on IFN-alpha generation in HCL patients pre- and post-therapy. Prior to treatment with 2-chloro-2'-deoxyadenosine (CdA), MNC from 24 HCL patients with active disease produced little or not IFN-alpha (geometric mean < 40 IU/ml) compared with controls (n = 140, geometric mean 1730 IU/ml, p < 0.0005). After treatment with CdA, IFN-alpha generation was studied in 16 patients, with a geometric mean value of 650 IU/ml (p < 0.0005 compared with pre-CdA levels). The severe depression of IFN-alpha generation improved progressively following CdA therapy-induced clinical remission. We propose that deficiency of IFN-alpha production may play a role in the susceptibility to intracellular infections of patients with active HCL.

Distinct characteristics of lymphokine-activated killer (LAK) cells derived from patients with B-cell chronic lymphocytic leukemia (B-CLL). A factor in B-CLL serum promotes natural killer cell-like LAK cell growth

We show that lymphokine-activated killer (LAK) cell precursors derived from patients with B-cell chronic lymphocytic leukemia (B-CLL) and cultured in the presence of recombinant interleukin-2 and normal human serum (NHS), develop into primarily NK cell-like (CD 57+) LAK cells, whereas identically prepared LAK cell precursors from normal subjects develop into mainly T cell-like (CD 3+, CD 8+) LAK cells. B-CLL LAK cells exhibited greater proliferative capacity than did normal LAK cells. When normal LAK cells were grown in B-CLL serum instead of NHS, their proliferation increased; NK cell levels also increased to those found in B-CLL LAK cells, suggesting that B-CLL serum contains a factor that promotes NK cell-like growth, LAK cells derived from normal or B-CLL patients demonstrated similar lytic activity toward K562 and Raji cells. Growth in B-CLL serum did not reduce their lytic potential. Thus, the altered phenotype and growth exhibited by B-CLL LAK cells and normal LAK cells grown in B-CLL serum does not lead to abnormalities in their cytolytic functions. We propose instead that the predominance of NK-like cells in B-CLL LAK cell populations and the presence of an NK cell-like growth factor in B-CLL serum reflect abnormalities related to NK cell-mediated B-cell regulation; ie, either inhibition of normal B-cell growth and/or growth stimulation of the leukemic clone in B-CLL.

Treatment of early chronic lymphocytic leukemia: intermittent chlorambucil versus observation

The effect of early therapy on the course of chronic lymphocytic leukemia (CLL) has not been established. Fifty-nine patients with indolent Rai stage I and II CLL were randomized to receive intermittent chlorambucil once a month or to receive no treatment. The two groups were comparable in entry characteristics. At 5 years from randomization there was no significant difference in survival between the two groups although the proportion of patients exhibiting active disease 5 years after randomization is 70 per cent in the untreated group and 55 per cent in the treated group. In this study, early treatment of CLL with intermittent chlorambucil did not result in a survival advantage for patients with indolent stage I and II CLL.

Hepatocellular carcinoma in a long-term survivor of acute lymphocytic leukemia

Hepatocellular carcinoma (HCC) occurred in a 28-year-old woman treated for acute lymphocytic leukemia (ALL) with methotrexate (MTX) and 6-mercaptopurine (6-MP), off all therapy for 15 years, who was also heterozygous for alpha-1 antitrypsin (alpha-1 AT) deficiency. MTXD is responsible for the development of HCC in this patient. The literature concerning the incidence of HCC in patients treated with MTX and 6-MP and in alpha-1 antitrypsin deficiencies is reviewed.

Chronic lymphocytic leukemia

This article reviews, for the internist, recent advances in our understanding of the immunology and clinical characteristics of chronic lymphocytic leukemia (CLL). The method of treatment based on clinical staging of CLL and as practiced in the authors' clinic is detailed. It also provides an outline of possible investigation and therapy that may be expected in the coming years.

Changes in body composition of cancer patients following combined nutritional support

The effects of combined nutritional support (parenteral, enteral, and oral) were measured in cancer patients unable to maintain normal alimentation. Changes in body composition were quantified by measurement of total body levels of nitrogen, potassium, water, and fat. The protein-calorie intake of the patients was also evaluated by dietary survey (4-day recall). Standard anthropometric and biochemical measurements for nutritional assessment were obtained for comparison. The dietary evaluation indicated that the dietary supplementation for all patients was more than adequate to meet their energy requirements. Almost all patients gained weight on the combined nutritional support regimens. Determination of body composition indicated that change in body weight was equal to the sum of the changes in body protein, total body water, and total body fat. The findings from the anthropometric nutrition indices (arm muscle circumference and triceps skinfold) were consistent with the results of the body composition study. Information on the nature of the tissue gained was obtained by comparison of body composition data with the ratio of protein:water:lean body mass for normal tissue. The mean gain of protein in the cancer patients was quite small (0.3-0.6 kg). The main change in body weight appeared to be the result of gains in body water and body fat. The total body nitrogen to potassium ratio served to define the extent of tissue anabolism following hyperalimentation. The ratio dropped in the cancer patients following hyperalimentation toward the value of the control subjects on ad libidum diets. The body compartment techniques described have demonstrated their usefulness in determining the effects of hyperalimentation on cancer patients.

Comparison of methods of estimating body fat in normal subjects and cancer patients

Total body fat can be indirectly estimated by the following noninvasive techniques: determination of lean body mass by measurement of body potassium or body water, and determination of density by underwater weighing or by skinfold measurements. The measurement of total body nitrogen by neutron activation provides another technique for estimating lean body mass and hence body fat. The nitrogen measurement can also be combined with the measurement of total body potassium in a two compartment model of the lean body mass from which another estimate of body fat can be derived. All of the above techniques are subject to various errors and are based on a number of assumptions, some of which are incompletely validated. These techniques were applied to a population of normal subjects and to a group of cancer patients. The advantages and disadvantages of each method are discussed in terms of their ability to estimate total body fat.

Biliary excretion of glycolipid in induced or inherited glucosylceramide lipidosis

Metabolically inert L-[1-14C]glucosylceramide is stored predominantly in the liver after intravenous administration to mice. The half-time of this glycolipid analogue in the liver is 3.5 days and its clearance occurs predominantly via the bile. Within the limited number of Gaucher specimens available for examination very high levels of glucosylceramide were found in the bile of one patient and in the liver of two patients with biliary obstruction. The question of a possible relationship between biliary excretion of glycolipid and the pathogenesis of Gaucher's disease will require further studies.

Compartmental body composition of cancer patients by measurement of total body nitrogen, potassium, and water

Quantitative measurement was made of body composition in patients with several forms of neoplastic disease. Total body nitrogen was determined by means of the prompt gamma neutron activation technique; total body potassium was measured with the use of a whole body counter. The mass and protein content of the muscle compartment and nonmuscle lean tissue were estimated by application of the technique of compartmental analysis. Total body water, determined simultaneously with the use of tritium label, provided a measure of lean body mass. From these data, the body fat can be inferred. The prompt gamma neutron activation and whole body counting techniques represent a considerable advance over the balance and radioisotope techniques used in earlier studies. The new techniques make possible sequential studies over prolonged periods of time with a considerable degree of accuracy. The loss of body weight by patients with solid tumors consisted primarily of the loss of muscle mass and body fat. Even in severe wasting, the patients appear to retain significant amounts of body fat. It is the skeletal muscle which is predominantly lost; the visceral life-supporting system is, to a considerable extent, spared. The nonmuscle tissue including the visceral fraction did not change in this study, and actually appeared to increase in size when comparison was made with the normal contrast population. The loss of total body water was slight in the cancer patients studied.

Compartmental body composition based on total-body nitrogen, potassium, and calcium

The techniques of prompt gamma neutron-activation analysis for the measurement of total-body nitrogen and whole-body counting for the measurement of total-body potassium were used to determine the mass of muscle and nonmuscle lean tissue and their protein content in 135 normal male and female subjects, 20-80 yr of age. Age-related changes in the size of the muscle and nonmuscle compartments and their protein content provide basic data for the investigation of protein metabolism in aging subjects and in individuals with various metabolic disorders, particularly wasting diseases such as cancer. Significant age-related changes in the size of various body compartments were noted. The loss of muscle mass and its protein content contrasts with the relative constancy of the nonmuscle lean tissue and suggests that skeletal muscle is particularly vulnerable to the aging process.

Bone marrow biopsy and clinical staging in chronic lymphocytic leukemia

We have examined the value of bone marrow biopsy histopathology in chronic lymphocytic leukemia (CLL). Cases involving a pattern of interstitial, nodular, or mixed interstitial plus nodular infiltration of lymphocytes were clinically similar and manifested a more benign course. These patterns correlated with the clinical Stages 0, I, and II. However, cases involving a pattern of diffuse or mixed diffuse plus nodular infiltration of lymphocytes formed a clinical group with a more rapidly progressive course and these patterns correlated with clinical Stages III and IV. Other measurements, e.g., calculation of the percentage of lymphatic tissue in the bone marrow biopsy specimen and of the ratio of the percentage of lymphatic tissue in marrow to the peripheral blood lymphocyte count, were found to be cumbersome to perform and of no additional value to clinical staging in prognosis indication in CLL.

A trypsin method for coating human red blood cells with plasma C3 and/or C4

Simple procedures are described for the production of human red blood cells coated with C3 and/or C4 by brief trypsin treatment of cell-plasma mixtures at room temperature. These cells ('EC3/4'-try) were unusual in that their agglutinability by anticomplement sera, as compared to sucrose-water RBC, was similar when tested by AutoAnalyzer but was markedly reduced in manual testing, especially against anti-C4. In spite of this, both 'EC34'-try and 'EC3'-try were found to be suitable indicator cells for determining the adequacy of anti-C3c and/or anti-C3d activities in reagents used for manual antiglobulin tests.

Mosaicism of red cell ABO type without recognizable cause

A 19-year old woman has been found whose red cells are a mixture of groups A1B (60%) and A1 (40%). Tests for many other genetic markers show no other evidence of mosaicism. She is not a twin and her parents and a sib are of normal ABO type. The cause of her dual population of red cells cannot be established, but may be an effect of somatic crossing over or mutation.

C3d antiglobulin haemagglutination of human red blood cells. A demonstration of two types of cell-bound C3d by means of typsin digestion

Washed human red blood cells from blood collected in EDTA were tested by Auto-Analyzer for a percentage of maximum antiglobulin haemagglutination (AH) using monospecific antisera to human C3d and C3c. The cells from normal persons were found to be agglutinated by anti-C3d but not by anti-C3c. To a fixed dilution of antiserum, the normal C3d AH values (X +/- 2 SD) were 34 +/- 19% for adult cells (n = 29) and 14 +/- 19% for cord cells (n = 19); the difference was significant (P less than 0.0001). By pretreatment of these cells with trypsin the C3d AH was either completely abolished or markedly reduced. Its difference between the adult and cord cells was eliminated as the observed values were 4 +/- 7% and 3 +/- 4% respectively (P = 0.15). The supernatant fluid of cell-trypsin mixture, treated with trypsin inhibitors, was found to be inhibitory to C3d AH but not to C3c AH. In contrast, the AH of C3d-coated red blood cells resulting from complement fixation in vivo (ie, cold agglutinin disease) or in vitro (eg, sucrose water reaction) was resistant to trypsin treatment. The difference between the trypsin-sensitive and trypsin-resistant cell-bound C3d is postulated to be at its attachment mechanism to the cell membranes. In addition, both the advantage and limitation of using trypsinised cells for C3d antiglobulin tests are demonstrated.

The low ionic strength reaction of human blood: relationship between the binding of serum immunoglobulin and complement of red blood cells and surface charge of the cells

Using the sucrose haemolysis reaction of Hartmann & Jenkins (1966) as a basic model, the low ionic strength reaction (LISR) of human blood was studied to determine: (1) serum Ig uptake by RBC with saline elution and 125I-IgG uptake, and (2) complement fixation (CF) to RBC with lysis of PNH cells and C3H/C4 antiglobulin haemagglutination (AH) of normal cells. The saline eluates were found to contain IgG and IgM with traces of IgA; their pH optima for the uptake by RBC were 6.0 +/- 0.5, 5.5 +/- 0.5 and c 5.0 respectively. The ratio of bound IgG to IgM was linearly related to the uptake pH. Both C4 AH and lysis were found to be optimum at pH 6.0--7.5, whereas the maximum C3 AH was at pH 6.0 +/- 0.5. The LISR performed at a constant pH (6.1 +/- 0.1) showed that an increasing concentration of neuraminidase (VCN) used in pretreatment of RBC was associated with a decrease in both IgG uptake and CF activity. A maximum VCN effect reduced the Ig uptake to c 20% of normal and abolished almost all the CF activity. An impaired LISR to various degrees was also observed with RBC pretreated with ficin, papain, bromelin, trypsin or protamine, and RBC from two individuals of En(a-) type. Preincubation of serum at LIS with and without RBC resulted in respectively a 'complete' and partial consumption of C in the fluid phase. The latter was not enhanced or inhibited by the addition of VCN-treated RBC for preincubation. A hypothesis is proposed suggesting that in the LSR the Ig uptake by RBC is an electrostatic interaction of the oppositely charged RBC and Ig and the CF to RBC results from C activation by the cell-bound IgG and IgM. In addition, a pH-dependent inactivation of the cell-bound C3 in the LISR is demonstrated.

Correlation of total body potassium and leukemic cell mass in patients with chronic lymphocytic leukemia

Total body leukemic mass in patients with chronic lymphocytic leukemia (CLL) was measured by quantitation of total body potassium (TBK) with a whole-body counter. In addition, the predicted normal total body potassium (Kp) for each patient was calculated from an empirically derived relationship involving height, weight, age, and sex. Both the absolute TBK and the relative excess of total body potassium (TBK/Kp) were related to the stage of disease. Patients in the early stages of CLL were found to have lower TBK and TBK Kp than patients in the late stages of disease. Both of these parameters increased with the successively advanced stages of the disease. The clinically monitored reduction of leukemic cell mass following therapy was accompanied by reductions in TBK and TBK/Kp. Data presented support the notion that TBK/Kp is a useful indicator of the total body leukemic mass. Furthermore, the results of these studies quantitatively validate the proposed clinical staging system for CLL. Quantitation of TBK by a whole-body counter is an accurate and noninvasive procedure and does not require administration of isotopes.

A comparative study of a BCNU containing 4-drug program versus MOPP versus 3-drug combinations in advanced Hodgkin's disease: a cooperative study by the Cancer and Leukemia Group B

A prospective randomized trial by CALGB examined the relative value of four chemotherapy regimens in 537 patients with stage III B and IV Hodgkin's disease. A new combination BOPP, derived by substitution of BCNU for nitrogen mustard in the MOPP regimen, was compared to MOPP and to two 3-drug regimens, derived by removing the procarbazine in BOPP (BOP) or removing the alkylating agent (OPP). The 4-drug programs gave significantly higher frequency of complete remissions (BOPP 67%, MOPP 63%) than the 3-drug regimens (BOP 40%, OPP 42%), and significantly longer duration of remission and survival. BOPP had a therapeutic activity equal to MOPP, and was accompanied by less toxicity. After 6 cycles of induction chemotherapy, responding patients, both CR and PR, were continued on maintenance chemotherapy for 3 years. No significant difference in relapse rate was demonstrated following maintenance treatment with either vinblastine, chlorambucil, or chlorambucil plus monthly vincristine + prednisone doses. Nor could a reinforcement phase late in the maintenance program be shown to influence the relapse rate. The median survival for all patients entered on the 4-drug programs was 5 years, while the median has not yet been reached at 6 years for those patients, who obtained CR.

Ripley-like anti-Rh associated with red blood cell-bound IgG aggregates

A commercial Rho (D) immune globulin after heating at 63 C became Ripley-like in the Rh-positive red blood cells coated with this heated globulin carried biologic activities similar to those of red blood cells coated with Ripley anti-CD serum. These coated red blood cells fixed complement and were agglutinated by all 20 sera containing rheumatoid factor (RF). The RF-Rh-hemagglutinations were more readily inhibited by heated than by unheated human IgG. The heated globulin had no such effect on Rh-negative red blood cells. Fractionation studies by Na2SO4 precipitation and/or Sephadex G-200 gel filtration revealed that heat-induced IgG aggregates in heated globulin were responsible for the biological activities. In contrast, these activities in Ripley serum were carried by IgG monomers. Another anti-CD serum (Heyman), tested in paralledl, was found to be indistinguishable from Ripley. a pooled RF serum, after multiple adsorptions with red blood cells coated with globulin, lost its agglutination activity to red blood cells coated with Ripley or Heyman serum.

A new form of Niemann-Pick disease characterised by temperature-labile sphingomyelinase

A new type (F) of Niemann-Pick disease characterised by childhood onset of splenomegaly, lack of neurological involvement, and diminished sphingomyelinase activity is described. The clinical presentation and heat-labile sphingomyelinase activity of this type F Niemann-Pick disease distinguishes it from other types of Niemann-Pick disease.

Guidelines for protocol studies in chronic lymphocytic leukemia

The treatment of any malignant disease requires appropriate criteria for diagnosis of the disease, appropriate methods for study, and techniques for evaluating response to treatment. In the past five years, there have been a number of proposed guidelines for studies in chronic lymphocytic leukemia (CLL). These submitted guidelines represent a further re-evaluation of concepts in CLL, and are currently in use by Cancer and Leukemia Group B for protocol studies. It is clear that revisions will continue to be make as new knowledge accumulates. In the meantime, these are submitted for publication for use by other investigators.

Comparison of daily versus intermittent chlorambucil and prednisone therapy in the treatment of patients with chronic lymphocytic leukemia

Ninety-six patients with stage III and stage IV chronic lymphocytic leukemia (CLL) were randomized into one of three treatment schedules. Prednisone was common to all three schedules and was given daily in an initial dosage of 0.8 mg/kg for the first 14 days, with successive halving of the daily dose on days 15 and 29 for a total 6-wk course. Prednisone was then given once a month at 0.8 mg/kg once a day for each of 7 consecutive days. Schedule I was prednisone plus chlorambucil (CLB) given as a once-a-month dose of 0.4-0.8 mg/kg; schedule II was both drugs, but the CLB was given as a daily dose of 0.08 mg/kg; schedule III was prednisone alone. Complete and partial remission (CR + PR) was 47% for schedule I, 38% for schedule II, and 11% for schedule III. Patients who responded (CR + PR) in each of the treatment schedules survived longer than the nonresponders. Complete remission was obtained in both CLB treatment schedules, but not with the prednisone alone regimen. Although overall survival was best in the intermittent CLB arm, there was no significant difference in survival time between the three treatment schedules. Toxicity was minimal in all three regimens. Augmentation of the intermittent monthly CLB, even to 1.5 and 2.0 mg/kg, was tolerated without undue marrow toxicity. About 22% of these patients either had diabetes mellitus at the time of entry on the study or manifested hyperglycemia during the course of treatment and observation.

Mediastinal irradiation for chronic lymphocytic leukemia

Thirty-one patients with chronic lymphocytic leukemia were treated with mediastinal radiation. In none of the patients was complete remission achieved; either partial remission or clinical improvement was achieved in 52 per cent, but the duration of response was short. The response rate was 77 per cent for the patients receiving a total radiation dose greater than 3,000 rads and 45 per cent for those receiving less than 3,000 rads. Severe life-threatening toxicity was noted in 11 patients and seven of these patients died; two patients died with progressive disease. Severe toxicity was manifested by one or more of the following: bone marrow aplasia, pancytopenia, gram-negative sepsis, generalized herpes zoster and severe esophagitis. Neither the total dose of radiation nor the dose per week correlated withe the severity of reaction or death.

Lymphocyte response to phytomitogens in iron deficiency

The lymphocytes of iron-deficient anemic and nonanemic patients have a decreased response to phytomitogens. Treatment of these patients with iron is followed by restoration of the lymphocyte response to normal. These findings indicate an important nonhemoglobin function of iron. Iron deficiency may be a factor in the production or potentiation of immune deficiency.

Factor IX deficiency in Gaucher disease. An in vitro phenomenon

Low factor IX coagulant activity was demonstrated in a patient with Gaucher disease before and after splenectomy. The same abnormality was found in seven of ten additional patients studied. Possibly, abnormally retained sphingolipids inactivate or interfere with the action of certain procoagulants.

Anti-Holley (hy): characterization of another example

Anti-Hy, in a serum of a black man who had received multiple blood transfusions, was found to be a warm reactive and polyclonal IgG antibody which did not bind complement in vitro. 51Cr-labeled Hy-positive red blood cells in the patient showed a shortened "2-component" survival curve. Hy antigens were not detected on leukocytes, platelets, or in the serum of Hy-positive individuals.

Clinical staging of chronic lymphocytic leukemia

A method of clinical staging of chronic lymphocytic leukemia (CLL) has been proposed which is based on the concept that CLL is a disease of progressive accumulation of nonfunctioning lymphocytes: stage O, bone marrow and blood lymphocytosis only; stage 1, lymphocytosis with enlarged nodes; stage II, lymphocytosis with enlarged spleen or liver or both; stage III, lymphocytosis with anemia; and stage IV:lymphocytosis with thrombocytopenia. Analysis of 125 patients. in the present series showed the following median survival times (in months) from diagnosis: stage 0, is greater than 150; stage I 101; stage II, 71; stage III, 19; stage IV, 19, The median survival for the entire series was 71 mo. The prognostic significance of the stage remained even after adjustment was made for age and sex. However, both sex and age were shown to be poor predictors of survival after adjustment for stage. The method of staging proved to be a reliable predictor of survival whether used at diagnosis or during the course of the disease. The proposed staging system was an equally accurate indicator for survival when applied to two other previously published studies of large series of patients