A double-blind, placebo-controlled study was carried out to see whether the synthetic E prostaglandin, misoprostol, would prevent gastric ulcer induced by non-steroidal anti-inflammatory drugs (NSAIDs). 420 patients with osteoarthritis and NSAID-associated abdominal pain were studied; they were receiving ibuprofen, piroxicam, or naproxen. Endoscopy was done at entry and after 1, 2, and 3 months of continuous treatment with 100 micrograms or 200 micrograms misoprostol or placebo, given four times daily with meals and at bedtime, concurrently with the NSAID. Abdominal pain was rated independently by patients and physicians. A treatment failure was defined as development of a gastric ulcer. Gastric ulcers (0.3 cm in diameter or greater) occurred less frequently (p less than 0.001) in both misoprostol treatment groups (5.6% 100 micrograms and 1.4% 200 micrograms) than in the placebo group (21.7%). The significant difference in ulcer formation between the placebo and the misoprostol treatment groups remained when comparisons were restricted to ulcers greater than 0.5 cm in diameter (12.3% placebo, 4.2% 100 micrograms misoprostol, and 0.7% 200 micrograms misoprostol). Mild to moderate, self-limiting diarrhoea was the most frequently reported adverse effect attributed to misoprostol. These results provide the first clear indication that NSAID-induced ulcers are preventable.

In laboratory animals, non-steroidal anti-inflammatory drugs (NSAIDs) are reported to inhibit diet-induced gallstone formation. To see if these drugs had a similar effect in man, 82 patients who had taken part in a comparison of ursodeoxycholic acid, placebo, and diet for prevention of gallstone recurrence were sent questionnaires about their use of NSAIDs during the period of the trial. 75 replied. After a mean follow-up of 33 (SEM 4) months none of the 12 regular users of NSAIDs had had gallstone recurrences, compared with 20 of the 63 who never or rarely used these drugs (p less than 0.02).

21 patients with AIDS and Kaposi's sarcoma were enrolled in an open therapeutic trial to determine the in vivo anti-retroviral activity of recombinant interferon-alpha (IFN-alpha). 8 (38%) showed a complete or partial anti-tumour response. The mean pretreatment CD4 count for the responders was 399 cells/microliter vs 154 cells/microliter for the non-responders. All 5 of the patients with more than 400 CD4 cells/microliter pretreatment showed a significant reduction in tumour, whereas none of the 7 patients with under 150 CD4 cells/microliter had any response. 5 of the 6 complete or partial responders with greater than 50 pg/ml of human immunodeficiency virus (HIV) p24 before IFN therapy showed a 75% or greater reduction by 12 weeks of therapy, with 3 patients having persistently negative HIV cultures. The anti-viral effects were also most pronounced in the patients with the highest CD4 counts. These data demonstrate the potential benefits, both anti-tumour and anti-retroviral, of treatment with IFN-alpha in the early stages of HIV infection and Kaposi's sarcoma.

134 Gambian children under 5 years of age with severe pneumonia (as defined by the World Health Organisation classification of acute respiratory infections) were given either oral co-trimoxazole for 5 days, or a single intramuscular dose of fortified procaine penicillin and 5 days of oral ampicillin. At 2 weeks, there was no significant difference in outcome between the two groups. Co-trimoxazole is much less expensive than ampicillin or procaine penicillin, requires only twice-daily administration, and can be given by health-care staff with little training. The results support the use of co-trimoxazole as the antibiotic of first choice in outpatient management of young children with pneumonia in developing countries.

Treatment with tamoxifen (TM), alone or in combination with cyclophosphamide, methotrexate, and fluorouracil (CMF), was used as an adjuvant to surgery in 433 patients with stage I, II, or III(T3a) breast cancer. Oestrogen receptors (ER) and progesterone (PgR) receptors were assayed in most cases. 308 premenopausal node-negative and postmenopausal node-negative or node-positive patients were randomised to receive TM, 30 mg daily for 2 years, or no further therapy. 125 premenopausal node-positive patients were randomised to receive either CMF for nine courses plus TM or CMF alone. After a median follow-up of 63 months TM significantly reduced the incidence of relapses and deaths compared with no therapy. A significant interaction between treatment effect and ER/PgR status was seen. Disease-free and overall survival were similar after treatment with CMF+ TM or CMF.

486 children born to HIV-positive mothers, 57 children infected by blood products, and 1 child for whom the personal history was not available were studied. Perinatal infection had a more varied clinical picture and a worse outcome compared with infection acquired later in childhood. Severe secondary infections, neurological disorders, and hepatitis (but not lymphoid interstitial pneumonia) were linked to a high mortality rate in perinatally infected children, in whom an early onset of symptoms was also a bad prognostic factor. Perinatal HIV infection occurred in 32.6% of children born to seropositive mothers, with a higher transmission rate in children born by vaginal delivery and then breast-fed. Preterm delivery and low birthweight seemed to be related to drug abuse during pregnancy, not to intrauterine HIV infection. Girls had a higher rate of perinatal infection and, of those infected, had an increased mortality.

271 children born to HIV-infected mothers in 8 European centres are being followed up from birth in a multicentre, collaborative study. By June, 1988, 45% had been followed for over 1 year: 10 had developed AIDS or AIDS-related complex, all by the age of 9 months, of whom 5 had died. 22 other children had symptoms or signs suggestive of HIV infection; of these, 12 had immunological abnormalities, 9 of whom were infected. 5 children had problems not related to HIV, including 3 neonatal deaths. The other 234 children are immunologically normal and clinically well. The median age of antibody loss was 10.3 months, although 1 did not lose antibody until over 18 months. None lost antibody and then became and remained seropositive. Of 100 children followed for more than 15 months, 19 had persistent antibody, and 5 were antibody-negative but presumed to be infected because of virus isolation or antigen detection; these 5 children were clinically and immunologically normal. The estimated vertical transmission rate was 24%.

The effect of cimetidine on survival was investigated in 181 patients with gastric cancer. Immediately after operation or the decision not to operate, the patients were randomised in double-blind fashion to placebo or cimetidine 400 mg twice daily for two years or until death, with review every three months. Median survival in the cimetidine group was 450 days (range 1-1826) and in the placebo group 316 days (1-1653). The relative survival rates (cimetidine/placebo) were 45%/28% at 1 year, 22%/13% at 2 years, 13%/7% at 3 years, 9%/3% at 4 years, and 2%/0% at 5 years. Survival in the cimetidine group was significantly longer than in the placebo group.

121 patients with acute pancreatitis thought to be due to gallstones were randomised to treatment with urgent endoscopic retrograde cholangiopancreatography (ERCP) and endoscopic sphincterotomy (ES) or with conventional treatment. They were stratified by predicted severity of the attack, according to the modified Glasgow system. ERCP was done within 72 h, and if common bileduct stones were identified, patients underwent ES immediately to extract the stones. There were fewer complications in the 59 patients who underwent ERCP +/- ES than among the 62 treated conventionally, the difference being confined to those whose attacks were predicted to be severe (6/25 ERCP +/- ES [1 death] compared with 17/28 conventional treatment [5 deaths]). Hospital stay was also shorter for patients with severe attacks who underwent ERCP +/- ES than for those who received conservative treatment (median 9.5 versus 17.0 days).

229 children with acute lymphoblastic leukaemia (ALL) and with clinical and laboratory features associated with a high risk of treatment failure entered a randomised study of three treatment regimens. Before 1981, such patients had a 3-year event-free survival (EFS) of 47%. Two intensive therapies, the Berlin-Frankfurt-Munster (BFM) 76/79 regimen and the New York (NY) regimen were compared with a control regimen that had achieved the best outcome in previous Trials. Data on 214 cases (93.4%) were analysed. The 3-year EFS was 78% for the BFM and NY regimens and 49% for the control regimen, a significant difference. The differences persisted after stratification by age at onset, sex, white blood cell count at diagnosis, and marrow blast morphology. Control patients were 2.7 times more likely to fail induction, to die, or to relapse than were patients on the intensive regimens.

Repeated intraperitoneal administration of therapeutic amounts of radiolabelled (iodine-131) murine monoclonal antibodies leads to the development of an immune response in the recipient, part of which is directed against the variable region (idiotype) of the administered antibody (anti-Id1 response). Human immunoglobulin purified from these patients inhibits binding of the original murine monoclonal antibody to its target tumour antigen and therefore represents an "internal image" of the tumour antigen. Furthermore, this study recorded the development of human antibodies that themselves bind to the tumour antigen, with a specificity identical or similar to that of the injected monoclonal antibody. These human antitumour antibodies are probably generated by way of the idiotypic network and confirm the existence of the idiotypic network. Accompanying this antitumour response the transient development of autoantibodies that react with connective tissue components of liver, kidney, spleen, and diaphragm was also observed.

Twelve short (more than two standard deviations below the mean height for age), prepubertal children (ten boys, two girls) who had a normal peak growth hormone (GH) response to provocative stimulation with clonidine (more than 10 ng/ml) were enrolled in a double-blind, placebo-controlled, crossover study of the effects of a single, nightly dose of clonidine (0.1 mg/m2 by mouth). The children's mean age was 7.2 years (range 3.6-10.5 years). The results of 6 months of clonidine therapy were compared with those of 6 months of placebo. Clonidine therapy resulted in no significant difference in height standard deviation score, growth velocity, bone age, 24 h integrated GH concentration, peak GH response to clonidine stimulation, levels of insulin-like growth factor 1, or predicted height by the RWT method. In contrast to other studies, this study shows no sustained increases in GH production or in improved growth velocity with long-term administration of a single daily dose of clonidine. Furthermore, this study demonstrates the need for well-designed, placebo-controlled trials in paediatrics.

Twenty remote-controlled insulin pumps (Siemens AG) were implanted into insulin-dependent type I diabetic patients for a one-year feasibility trial in four centres. The total observation time was 18.2 patient-years. Three pumps had to be prematurely removed after 101, 141, and 236 days. Patients self-monitored blood glucose levels with a mean of 5.5 (range 1-17) measurements per day. 62.9% of these measurements were in the range 3.33-8.88 mmol/l. 3.25 glucose measurements per patient-month were in the hypoglycaemic range (lower than 2.78 mmol/l) and 2.6 episodes of hypoglycaemia with symptoms were reported per patient-month, of which 0.22 per patient-year required medical attention. The median HbA1c level was 7.6% at baseline (10-90% centile = 5.9-9.1%) and 7.0% at the end of the trial (10-90% centile = 5.7-8.3%) (p less than 0.05). Despite some technical and clinical problems, the pump, when used with a stable insulin preparation, was an effective means of treating insulin-dependent patients.