958 provide controls-ie, verified, drug-free AIDS cases-and because it does not quantify drug use and ignores zidovudine use altogether. Among the 365 HIV-positive men surveyed, 88% reported consumption of nitrites and 80% the use of other so-called illicit recreational drugs. Thus, assuming no linkage, 98% have used drugs, because only about 2% did not report illicit drug use and 70% have used at least two drugs. In- fact 100% could have used drugs, especially since illicit drug use is notoriously underreported and since Schechter et al did not verify drug use by any tests. Moreover, another study by Schechter’s group reports that 100% of 87 AIDS patients from the Vancouver cohort had used nitrites.3 Thus Schechter’s data confirm rather than refute the drug-AIDS hypothesis. Moreover, drug use information was obtained on an ever-versus- never basis, rather than being quantified. But with drugs, the dose is the poison. 1 year of smoking or drinking may not cause lung cancer or liver cirrhosis, but 10-20 years’ may do so. Indeed, Schechter’s data also confirm the quantitative aspect of the drug hypothesis if one considers that antibody against HIV is a marker for drug consumption, as they do ("risk behaviours are known to corelate to HIV-1 infection . Since it takes about 1000 drug-promoted sexual contacts to be infected with HIV, HIV antibody-positive individuals have been exposed to the drug equivalent of 1000 contacts more than those who are HIV-negative. Given a choice between drugs and antibodies, it is more rational to conclude that drug intake has an "integral role in CD4 depletion and AIDS" than antibodies that neutralise HIV.1 1 The T-cell losses that Schechter et al attribute to HIV can also be explained by zidovudine, which is prescribed as AIDS prophylaxis and therapy to HIV-positive, but not to HIV-negative individuals. Originally developed for chemotherapy to kill all dividing cells, zidovudine is especially toxic to the rapidly proliferating bone marrow, the source of T -cells. 1,4,5 Thus treatment with zidovudine and a higher lifetime dose of recreational drugs in positive than in negative subjects explain the decline of T-cells in HIV-positive individuals. In the same issue (p 686) you review a report by Ascher et al who also claim that HIV causes AID S independent of drugs, on the basis of surveys of mostly homosexual men from San Francisco .6 HIV seropositive/no-drug users are claimed to lose T-cells at the same rate as drug users. However, there were no HIV-positive/no-drug users in Ascher’s study, although these workers imply that there were "homosexual/bisexual men reporting none [drug use]". This statement is documented as follows. According to table i all HIV-positives were homosexuals. All heterosexuals were HIV- negative, apart from 1 who was a drug-addict (Ascher MS, personal communication). According to table 11 all the homosexual men were either "heavy" or "light" nitrite users-namely, 144 plus 668 (table 11) out of 812 (text). An unreported percentage of these man had also used other illicit recreational drugs, such as cocaine and amphetamines, in addition to zidovudine.1 Therefore, all seropositive men in the study used drugs. However, the subsequent figure purports to give data on T-cell losses ofthree seropositive groups, with "no drug use", "moderate drug use", and "heavy drug use", respectively. Three curves on that figure correspond to these three groups. Yet based on tables i and 11 the category "seropositive/ no drug use" is an empty set representing nobody. The curve, clearly labelled "seropositive-no drug use", is therefore a fabrication. Furthermore, Ascher et al directly confirm my drug- AIDS hypothesis, because all their AIDS cases were drug users, and because heavy users were twice as likely to develop AIDS as light users (table II). To refute my drug hypothesis Schechter or Ascher would have to produce a controlled study showing, that over a period of up to 10 years HIV-positive individuals who use recreational drugs or zidovudine or both have the same AIDS risks as positives who do not do so. The 10 year period is claimed by proponents of the HIV hypothesis to be the time needed for HIV to cause AIDS. Alternatively, they could show that HIV-free individuals who have used drugs for 10 years never get AIDS-defining diseases. Since Schechter et al do not "rule out a role for cofactors" for HIV, we may have more common ground than their report suggests. The burden of proof is on them to find these elusive cofactors, and if they are found, to show that they depend on antibodies against HIV to cause AIDS. Clearly HIV itself cannot contribute much to the development of AIDS, because it infects only 1 in 1000 T-cells.1 Further, HIV contributes nothing to the over 3000 HIV-free AIDS cases that have been recorded.1,8 It is for this reason that I hope the drug hypothesis will become a hindrance to the physiologically (zidovudine) and psychologically toxic (positive AIDS test) public health initiatives of the unproven HIV-AIDS hypothesis. Department of Molecular and Cell Biology, Stanley Hall, University of California, Berkeley, California 94720, USA PETER DUESBERG 1. Duesberg PH. AIDS acquired by drug consumption and other noncontagious risk factors. Pharmacol Ther 1992; 55: 201-77. 2. Duesberg PH. The role of drugs in the origin of AIDS. Biomed Pharmacother 1992; 46: 3-15. 3. Archibald CP, Schechter MT, Le TN, Craib KJP, Montaner JSG, O’Shaughnessy MV. Evidence for a sexually transmitted cofactor for AIDS-related Kaposi’s sarcoma in a cohort of homosexual men. Epidemiology 1992; 3: 203-09. 4. Kolata G. Imminent marketing of AZT raises problems: marrow suppression hampers AZT use in AIDS. Science 1987; 235: 1462-63. 5. Hamilton JD, Hartigan PM, Simerkoff MS, et al. A controlled trial of early versus late treatment with zidovudine in symptomatic human immunodeficiency virus infection. N Engl J Med 1992; 326: 437-43. 6. Ascher MS, Sheppard HW, Winkelstein W Jr, Vittinghoff E. Does drug use cause AIDS? Nature 1993; 362: 103-04. 7. Lang W, Osmond D, Samuel M, Moss A, Schrager L, Winkelstein W Jr. Population-based estimates of zidovudine and metosol pentamidine use in San Francisco. 1987-1989. J AIDS 1991; 4: 713-16. 8. Duesberg P. HIV-free AIDS reports. Science 1992; 257: 1848. Rapid decline of CD4+ cells after IFN&agr; treatment in HIV-1 infection SiR,&mdash;Interferon (IFN) is increasingly used for chronic hepatitis C virus (HCV) induced liver disease and its use is also envisaged in HIV-1 I infection.’ We report three symptom-free HIV-1 infected individuals with HCV related chronic active hepatitis (CAH) who had a rapid, profound decline of CD4 cells after IFNot. The patients’ details are in the table. Case 1 was male, 32, HIV-1 seropositive since 1984, and had HCV induced CAH in October, 1990. HLA haplotype is Al, 2; B8, 44; DR3, 7; DQ2,3. Recombinant IFNr:tza (Roferon) 6 MU thrice weekly was started in February, 1991. When we saw him in November, 1991, he had oral candidosis. IFN was stopped and zidovudine 500 mg daily was started. ALT is normal, and liver histology shows chronic persistent hepatitis. The patient has hairy leukoplakia. The second case was male, 28, a former intravenous drug user (IVDU), HIV-1 seropositive since 1985, and had HCV induced CAH in February, 1991. HLA haplotype is Al,19; B8, 35; DR3, 6; DQ1,4. Lymphoblastoid IFNa (Wellferon) 6 MU was started thrice weekly in April, 1991. When we first saw him, 10 months later in February, 1992, ALT was raised and CD4 cells were reduced. He had oropharyngeal candidosis and hairy leukoplakia. IFN was withdrawn and zidovudine 500 mg daily started. In October, 1992, despite cotrimoxazole prophylaxis, the patient had Pneumocystis carinii pneumonia. Liver histology still shows CAH. DETAILS OF PATIENTS *As% of total T cells ALT= alanine aminotransferase, N= normal (<22 U/L).