Patients with CLL seldom die because of a high white cell count, but there are plenty of other fatal characteristics of the disease. Immunodeficiency. Before the Second World War, Wintrobe recognized that patients were particularly sensitive to
infection (Wintrobe & Hasenbush, 1939) and we now know that most patients will develop low levels of immunoglobulin if cases are followed for long enough. Hypogammaglobulinaemia was first recognized as a clinical entity by Bruton (1952), when electrophoresis of serum revealed the surprising absence of the gamma fraction in an otherwise normal pattern. Immunological studies showed a complete absence of antibodies and isohaemagglutinins. Odd case reports of agammaglobulinaemia in CLL began appearing shortly afterwards (Brem & Morton, 1955; Jim & Reinhard, 1956). Jim (1957) found that 17 out of 50 patients had hypogammaglobulinaemia. In a more comprehensive study, Cone & Uhr (1964) found deficiencies in all classes of serum immunoglobulins and failure to produce an antibody response to phi x 174, a primary antigen, or diphtheria toxoid, a secondary antigen. There was also a failure to sensitize to dinitrofluorobenzene, although most patients did produce a delayed hypersensitivity response to recall antigens. Later studies correlated the more severe falls in serum immunoglobulins with more advanced disease (Fiddes et al, 1972) and demonstrated that it was possible to generate an immune response to a new antigen (phi x 174) in early stage patients if sufficient inoculations were given (Hamblin et al, 1975). Really profound immunodeficiency had to await the arrival of the purine analogues as a popular treatment (O'Brien et al, 1993). Why the mmunodeficiency of CLL is so much worse than that of other lymphoid tumours is one of the unsolved mysteries of the disease.
Winifred Ashby (1879-1975) moved from London to Chicago at the age of 14. She carried out her pioneering work into the life span of red cells at the Mayo Clinic between 1917 and 1921. Her technique (Ashby, 1919) involved the transfusion of red cells that were compatible with, but serologically distinct from, those of the recipient and then tracking their survival by differential agglutination. Berlin (1951) used this technique in nine patients with CLL. All had a shortened red cell survival, even though only one had a reticulocytosis. This was probably the first demonstration that the anaemia of CLL might be haemolytic in nature.
It was shortly after Ehrlich (Ehrlich & Morgenroth, 1901) published the concept of `horror autotoxicus', the idea that the body would not make an antibody that destroyed its own tissues, that Donath & Landsteiner (1904) described an antibody that did just that. Shortly afterwards, Fernand Widal (of typhoid fame) was probably the first to recognize acquired haemolytic anaemia with red cell agglutination (Widal et al, 1908). Thirty years later, in Boston, Dameshek & Schwartz (1938) stressed the importance of `haemolysins' in the commonest type of acquired haemolytic anaemia. It was not quite clear what these `haemolysins' were until the development of the indirect anti-globulin test by Robin Coombs (Coombs et al, 1945) and its application in haemolytic anaemia (Boorman et al, 1946). Wasserman (not of syphilis fame) found haemolytic anaemia to be present in nine out 58 consecutive patients with CLL; five out of seven patients tested had a positive Coombs' test (Wasserman et al, 1955). A series of studies suggested that autoimmune haemolytic anaemia (AIHA) occurs at some time in the course of CLL in between 10% and 26% of cases (Wasserman et al, 1955; Pisciotta & Hirschboeck, 1957; Beickert, 1959; Dameshek & Schwartz, 1959; Troup et al, 1960; Videbak, 1962).
It is often forgotten that the `I' in ITP (immune thrombocytopenic purpura) originally stood for `idiopathic' and not `immune'. William Dameshek, although a giant of haematology almost without equal, was clearly wrong in his championing of `hypersplenism' as the cause of thrombocytopenia. The spleen was supposed to produce a sort of miasma that inhibited the bone marrow. Harrington et al (1951) first demonstrated that the plasma of patients with chronic ITP transfused into a normal recipient (himself) would produce thrombocytopenia. (Oh what experiments you could do in the world before viruses!) Later, Shulman et al (1965) showed that this plasma factor was present in the 7S gamma globulin fraction and was absorbed by human platelets.
Thrombocytopenia is quite common in CLL. Minot and Buckman (1925) found it in half their patients at presentation and in virtually all those patients whose white cell count rose above 175k. Harrington & Arimura (1961) reported seven cases of autoimmune thrombocytopenia occurring in CLL. Dameshek reported five more (Ebbe et al, 1962), but because of the unsatisfactory nature of platelet antibody tests the true prevalence of ITP in CLL is unknown. An increase in bone marrow megakaryocytes remains the surest touchstone.
Reporting autoimmunity in CLL was then a popular sport. Immune neutropenia (Killman, 1959), pure red cell aplasia (Abeloff & Waterbury, 1974), Sjogren's syndrome (Lehner-
Netsch et al, 1969), nephrotic syndrome (Dathan et al, 1974), bullous pemphigoid (Cuni et al, 1974), Graves's disease (Haubenstock & Zalusky, 1985), systemic lupus
erythematosus, rheumatoid arthritis, ulcerative colitis, allergic vasculitis and pernicious anaemia (Miller, 1962; Dameshek, 1967) have all been associated with CLL. The fact that CD5-positive B cells secrete antibodies that can be made to react with DNA and IgG (Sthoeger et al, 1989) has encouraged speculation concerning the origin of autoimmune disease in CLL. In fact, the clinically important autoantibodies are produced by the residual normal immune system, not by the tumour cells. Most of the associations between CLL and autoimmune diseases occur by chance. Only AIHA and ITP are more common than in an age-matched control population (Hamblin et al, 1986).
The high incidence of AIHA in fludarabine-treated patients (Myint et al, 1995) suggests that autoimmunity is a consequence of the severe immunodeficiency that occurs in CLL and especially the AIDS/like syndrome that may follow treatment with fludarabine.
Richter (1928) described an aggressive lymphoma occurring in a patient with CLL and gave his name to a phenomenon that occurs in up to 3% of patients. Histologically, the tumour is a diffuse large-cell lymphoma (Trump et al, 1980). Modern techniques have demonstrated that in roughly half the cases the second lymphoid tumour is clonally unrelated to the first (Miyamura et al, 1990; Kruger et al, 1993). In prolymphocytic transformation of CLL (Enno et al, 1979), the cell markers remain CLL like. Although of grave consequence to the patient, he does not develop PLL. Transformation to acute lymphoblastic leukaemia has been reported (Brouet et al, 1973). Such was the confusion at the time over what CLL was, I doubt whether this or any of the subsequent cases were correctly assigned.