1107 Recovery of platelet count and haemoglobin in patient with apparent treatment-resistant TTP. falsely lead readers in to believing that the treatment may have produced a favourable outcome. The role of IVIG, vincristine, and other treatments in TTP will only be clarified by randomised controlled trials, even though this will be difficult for such a rare condition. Department of Haematology, Royal Hallamshire Hospital, Sheffield S10 2JF, UK MIKE MAKRIS MIKE GREAVES 1. Wong P, Itoh K, Yoshida S. Treatment of thrombotic thrombocytopenic purpura with intravenous immunoglobulin. N Engl J Med 1986; 314: 385-86. 2. Messmore HL, Yeshwant C, Remlinger K, Seghatchian MJ. Intravenous gamma globulin in refractory thrombotic thrombocytopenic purpura. Thromb Haemost 1985; 54: 127. 3. Viero P, Cortelazzo S, Buelli M, et al. Thrombotic thrombocytopenic purpura and high-dose immunoglobulin treatment. Ann Intern Med 1986; 104: 282. 4. Finn NG, Wang JC, Hong KJ. High dose intravenous gamma-immunoglobulin infusion in the treatment of thrombotic thromobytopenic purpura. Arch Intern Med 1987; 147: 2165-67. Absence of Epstein-Barr virus nuclear antigen 2 in tumour cells of Hodgkin’s disease SIR,-Dr Pallesen and colleagues (Feb 9, p 320) demonstrate with immunohistochemistry the expression of Epstein-Barr virus (EBV) encoded latent gene product-latent membrane protein (LMP)-in a high proportion of patients with Hodgkin’s disease. With earlier reports of the presence of EBV genome by molecular methods and by in-situ hybridisation with radioactive probes,l,2 and our recent finding of EBV mRNA transcripts exclusively in phenotypically characteristic Reed-Stemberg cells by the use of non-isotopic probes,3 EBV seems to be directly implicated in the pathogenesis of Hodgkin’s disease. Wu et al by in-situ hybridisation with riboprobes, also provide evidence for latent gene transcription (EBER 1) of EBV in Reed-Stemberg cells. Appropriate levels of transcription and translation of the EBV latent genes have major implications in the EBV-induced immortalisation and growth of B lymphocytes and EBV-associated lymphomagenesis. Such studies might well eventually result not only in elucidation of the origin of the pathogenic cells of Hodgkin’s disease, but also affect the diagnosis, treatment, and, possibly, prevention of the disease. However, the lack of EBV nuclear antigen 2 (EBNA2) in the presence of LMP, as Pallesen et al found by the use of monoclonal antibodies, is intriguing and raises several questions. We have shown that Bam HI-W fragment, which encodes for the EBNAs/ is actively transcribed in about 30% of the nodular sclerosis (NS) and mixed cellularity (MC) subtypes of Hodgkin’s disease, but rarely in B-cell lymphomas unrelated to HIV.3 Strains of EBV with deletions of EBNA2 gene are deficient in their B-lymphocyte transformation properties.6,7 The role of EBNA2 is further emphasised by transactivation of LMP15 and because LMP is expressed on cell surfaces only when infected by immortalising strains of EBV.7 Rogers et a 15 have suggested that down-regulation of the EBNA2 gene could occur by formation of a leftward-rightward transcript duplex, but only during the lytic cycle of the virus-a situation not applicable to Reed-Stemberg cells. Therefore one wonders if the difference between Hodgkin’s disease (LMP+, EBNA2-) and other EBV-associated disorders shown by Pallesen et al is a de facto molecular event. If the findings reflect a true restrictive pattern in Hodgkin’s disease, then the possibility of non-functional mRNA transcripts or a post-translational defect needs to be considered. An alternative explanation is a structurally different EBNA2 not recognised by the PE2 antibody used by Pallesen et al. Lack of EBNA2 in Hodgkin’s disease should be confirmed by specific cDNA probes and anti-EBNA2 monoclonal antibodies against different epitopes, in order for the restrictive pattern to be applied in the physiopathology of EBV-associated Hodgkin’s disease. We also think that the low rate of EBV detection in lymphocyte-predominant Hodgkin’s disease noticed by us for EBV mRNA transcripts (0/6) and by Pallesen et al for LMP (1/10), in comparison to other subtypes (NS and MC), is important because this type of Hodgkin’s disease is being considered as a non-Hodgkin B-cell lymphoma in evolution.8 Although the origin of Reed- Sternberg cells is still unknown, Pallesen and colleagues’ findings and those of others, 1-4 might well implicate the participation of the preferred EBV target-the B lymphocyte-in the production of some or most cases of Hodgkin’s disease. Department of Anatomical Pathology, INSERM U 326, University Paul Sabatier (CHU Purpan), 31059 Toulouse, France PIERRE BROUSSET SHASHIKANT CHITTAL LI ROYA GEORGES DELSOL 1. Weiss LM, Movahed LA, Wamke RA, Sklar J. Detection of Epstein-Barr virus genomes m Reed-Sternberg cells of Hodgkin’s disease. N Engl J Med 1989; 320: 502-06. 2. Anagnostopoulos I, Herbst H, Niedobitek G, Stein H. Demonstration of monoclonal EBV genomes in Hodgkin’s disease and Kil positive anaplastic large cell lymphoma by combined Southern blot and in situ hybridization. Blood 1989; 74: 810-16. 3. Brousset P, Chittal S, Schlaifer D, et al. Detection of Epstein-Barr messenger RNA in Reed-Sternberg cells of Hodgkin’s disease by in situ hybridization with biotinylated probes on specially processed modified acetone methyl benzoate xylene (ModAMeX) sections. Blood (in press). 4. Wu TC, Mann RB, Charache P, et al. Detection of EBV gene expression in Reed-Stemberg cells of Hodgkin’s disease. Int J Cancer 1990; 46: 801-04. 5. Rogers RP, Speck SH. Bidirectional transcription of Epstein-Barr virus major internal repeat. J Virol 1990; 64: 2426-29. 6. Wang F, Tsang SF, Kurilla MG, Cohen JI, Kieff E. Epstein-Barr virus nuclear antigen 2 transactivates latent membrane protein LMP1. J Virol 1990; 64: 3407-16. 7. Cordier M, Calender A, Billaud M, et al. Stable transinfection of Epstein-Barr (EBV) nuclear antigen 2 in lymphoma cells containing the EBV P3HR1 genome induces expression of B-cell activation molecules CD21 and CD23. J Virol 1990; 64: 1002-13. 8. Chittal S, Alard C, Rossi JF, et al. Further phenotypic evidence that nodular, lymphocyte-predominant Hodgkin’s disease is a large B-cell lymphoma in evolution. Am J Surg Pathol 1990; 14: 1024-35. Monitoring CD4 counts in HIV infection SIR,—Dr Phillips and colleagues’report (Feb 16, p 389) on CD4 counts and development of AIDS among patients with haemophilia resembles our experience with homosexuals. We found a mean CD4 count of 60/uJ when AIDS develops. However, the rate of decline of CD4 count was variable and steeper in our group. CD4 counts among patients in whom AIDS developed ranged from 10 to 210/µl. The repeated antigenic stimulation of HIV-infected CD4 cells in homosexuals probably accounts for this difference. Certain viruses (ie, herpes simplex, hepatitis B, and cytomegalovirus) enhance replication of HIV, leading to further depletion of CD4 cells and progression to AIDS. If lymphocyte marker counts are to be used to monitor progression to AIDS, physicians will need to understand the basis for interpreting them. Factors that can influence the CD4 count are: acute intercurrent infection; diurnal variation (peaks at 2300 hours and troughs at 1100 hours); stress and exertion; temperature of specimen in transit; and delay in analysis (if samples are prepared within 24 h of venesection and transported in anticoagulant at an ambient temperature of 22°C, lymphocyte subpopulation results are relatively stable1).