615 limiting step in the degradation of haem to bilirubin. Bilirubin is a potent antioxidant. Haem oxygenase is active only in the presence of oxygen and so would be activated during the ensuing reperfusion events; this accords with in-vitro data showing that different stress proteins are released in hypoxia and following reperfusion.7 Thus in the hypoxic joint the 32 kD protein could have an important scavenging role. The anti-rheumatic drug auranofin induces synthesis of the 32 kD protein in vitro 8 This is a direct effect, dependent on auranofin crossing the cell membrane; nevertheless, both D-penicillamine and gold(I) thiomalate can induce the formation of ROS and might indirectly induce synthesis of the 32 kD protein in vivo. A target of hypoxic/reperfusion injury is the endothelial cell. Angiogenesis, which follows myocardial infarction, is also a feature of rheumatoid synovia.1 In the presence of tumour necrosis factor alpha, which is found in rheumatoid synovial fluid, the 28 kD stress protein is phosphorylated and can activate endothelial cells,9 possibly initiating proliferation or the release of angiogenic factors. Cytoskeletal derangements occur in hypoxic conditions and the 28 kD protein associates with actin microfilaments (B. Geiger, personal communication), causing changes in morphology and perhaps in vascular permeability. The 28 kD protein may be another focus for drug development. Other cytoskeletal components may have a role in vascular permeability. The intermediate filament vimentin is much increased in synovial cells and is very sensitive to oxidative damage. ROS, like heat shock, cause collapse of the vimentin cytoskeleton around the nucleus,1O where it becomes associated with a 70 kD stress protein. Disodium aurothiomalate at 01 mmol/1 induces a similar rearrangement of intermediate filaments in fibrolasts (K. R. Rogers, unpublished). The highly conserved stress proteins control the basic biology of cellular interactions, and the complex nature of the biochemical processes involved deserves careful exploration. Moreover, apparently diverse therapeutic approaches to suppress rheumatoid disease may have in common the ability to induce stress protein synthesis. Bone and Joint Research Unit, ARC Building, London Hospital Medical College, London E1 2AD, UK V. R. WINROW D. R. BLAKE 1. Stevens CR, Williams RB, Farrell AJ, Blake DR. Hypoxia and inflammatory synovitis: observations and speculation. Ann Rheum Dis 1991; 50: 124-32. 2. Allen RE, Blake DR, Nazhat NB, Jones P. Superoxide radical generation by inflamed human synovium after hypoxia. Lancet 1989; ii: 282-83. 3. Ananthan J, Goldberg AL, Voellmy R. Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat shock genes. Science 1986; 232: 522-24. 4. Winrow VR, Mojdehi G, Mapp PI, Rampton DS, Blake DR. Immunohistological localisation of stress proteins in inflammatory tissue. In: Burdon R, Rice-Evans C, Blake D, Winrow V, eds. Stress proteins in inflammation. London: Richelieu Press, 1990: 237-51. 5.de Graeff-Meeder ER, Voorhorst M, van Eden W, et al. Antibodies to the mycobacterial 65-kd heat-shock protein are reactive with synovial tissue of adjuvant arthritic rats and patients with rheumatoid arthritis and osteoarthritis. Am J Pathol 1990; 137: 1013-17. 6. Keyse SM, Tyrrell RM. Heme oxygenase is the major 32-KDa stress protein induced in human skin fibroblasts by UVA radiation, hydrogen peroxide, and sodium arsenite. Proc Natl Acad Sci USA 1989; 86: 99-103. 7. Sciandra JJ, Subjeck JR, Hughes CS. Induction of glucose-regulated proteins after anaerobic exposure and of heat-shock proteins after reoxygenation. Proc Natl Acad Sci USA 1984; 81: 4843-47. 8. Caltabiano MM, Poste G, Greig RG. Induction of the 32-kD human stress protein by auranofin and related triethylphosphine gold analogs. Biochem Pharmacol 1988; 37: 4089-93. 9. Darbon JM, Issandou M, Tournier JF, Bayard F. The respective 27kDa and 28kDa protein kinase C substrates in vascular endothelial and MCF-7 cells are most probably heat shock proteins. Biochem Biophys Res Commun 1990; 168: 527-36. 10. Rogers KR, Morris CJ, Blake DR. Cytoskeletal rearrangment by oxidative stress. Int J Tiss React 1989; 9: 309-14. Can disposables make matters worse? SiR,—Dr Rudin and colleagues’ letter on HIV, hepatitis, and measles in Romanian children (Dec 22/29, p 1592-93) leaves us wavering between sadness and despair. The situation is terrible: 12 of 20 orphans in hospital in Pascani, north-eastern Romania were HIV positive and most had evidence of exposure to hepatitis B virus. Rudin et al blame the "lack of disposable needles and medical supplies and the policy of transfusion of untested blood" for this spread of bloodborne diseases among children, and they suggest an international effort to supply disposable needles and syringes. But the situation is more complex than this: a closer look reveals that this "solution" could have an effect opposite to that which might be expected. In Romania the distribution of disposable injecting equipment was a top priority of donor countries and agencies and some hospitals were quickly well supplied with such items, while in others people fought hard to obtain them. The disposable materials were enthusiastically adopted, but the lack of safe disposal points and inadequate education of the public resulted in the presence of large numbers of used, uncleaned needles and syringes in the environment. Moreover, the abundance of disposables favoured the spread of parenteral treatment outside hospitals. The discontinuity in availability of disposable needles and syringes meant frequent switching to the former practice of making do with reusable materials, but now there was antipathy to the extra work of cleaning and sterilisation. Former routines of needle and syringe sterilisation were sometimes forgotten; the media cast doubt on the efficacy of these procedures and as a result the public have tended to reject old-fashioned glass syringes. Thus pressure to re-use non-re-usable materials is at work. When people’s behaviour and the social infrastructure are the culprits, as here, a measure such as the distribution of needles and syringes should be supplemented by a programme of re-education to ensure that every type of needle and syringe is used properly and that instructions are adhered to. Whenever the focus of health intervention is an object, rather than people with human feelings, an apparently simple measure such as distributing needles and syringes will have the capacity to create explosive epidemics. As has been argued in another context, development aid may add only yet more disturbance when the ecological foundation of health is neglected.’ 1 This letter could not have been written in the Ceausescu era, when it was an offence punishable by imprisonment to receive a western medical journal? Today we are fortunate to have complimentary subscriptions offered by journals, such as The Lancet and its readers3 or received by exchange agreements between our local medical journal Revista Medico-Chirurgicala and similar western publications. To cite a British opinion, "there are resources that may rescue Romania from total disaster: the intelligence, inventiveness, and wit of its people".2 This feedback is part of that package. Pulmonary Diseases Clinic, Institute of Medicine and Pharmacy, lasi 6600, Romania T. MIHAESCU L. VERES 1. Rijpma S, Barnes J, Meegan MK, Taylor CE. The "demographic trap". Lancet 1991; 337: 50-51. 2. Anon. Medicine in Romania. Br Med J 1990; 300: 699. 3. Heley MM. Journals for Romania. Lancet 1990; 336: 1013. Optic neuritis associated with dideoxyinosine SIR,-2’,3’-dideoxyinosine (ddl) is being studied in patients with AIDS or AIDS-related complex who do not tolerate or who do not respond to zidovudine. The major adverse reactions to ddl are peripheral neuropathy and pancreatitis.2 We report a case of optic neuritis 6 weeks after the introduction of ddI. Although the relation between ddl and optic neuritis cannot be asserted, the stabilisation of visual loss after interruption of ddl and its deterioration when this therapy was resumed are consistent with a toxic role of the drug. A 40-year-old homosexual man with HIV infection (CDC class IVC2 [hairy leukoplakia of the tongue]) had been treated with zidovudine 1200 mg daily since 1988 because of a CD4 count of 190/ul. In June, 1990, oesophageal candidiasis developed and the CD4 count was only 70/tit. 1 month later a herpes zoster eruption was treated with acyclovir. The CD4 count was 20/ul. Zidovudine was discontinued, and in October, 1990, ddl was introduced at a dose of 8 mg/kg daily, associated with fluconazole 50 mg per day