Previous studies have suggested that patients with Barrett's oesophagus may be at increased risk of colorectal neoplasia, though the association is disputed. In a multicentre prospective study we compared the prevalence of colorectal adenomas in patients with Barrett's oesophagus and controls. Barrett's oesophagus patients (n = 104) had histological confirmation of columnar epithelium extending more than 3 cm above the gastro-oesophageal junction. The 537 controls were patients with symptoms suggesting irritable bowel syndrome. No participant had a personal history of colonic neoplasm. Each patient underwent colonoscopy. Histologically proven adenomas were found in 26 Barrett's patients (25%) and 75 controls (14%). Three colorectal cancers were discovered in each group. The prevalence of adenomas was greater in the Barrett's oesophagus group than in the control group (p < 0.01) but the relation became non-significant after adjustment for age and sex and control for other known risk factors by a logistic regression model (odds ratio 1.4 [0.7-2.7]). The relative risk of adenoma was significantly higher in patients older than 59 than in younger patients (2.2 [1.3-3.5]) and in men than in women (3.4 [2.0-5.7]). Other factors contributing significantly to the risk of adenoma were a family history of colorectal cancer (2.3 [1.1-4.8]), rectal bleeding (2.1 [1.1-3.9]), previous colonic investigation (0.3 [0.1-0.7]), and complete as opposed to partial colonoscopy (6.4 [0.8-48.3]). We conclude that Barrett's oesophagus is not an independent risk factor for colorectal neoplasia and, therefore, is not, in itself an indication for colorectal screening.

The major motor disturbances in Parkinson's disease are thought to be caused by overactivity of the internal segment of the globus pallidus (GPi), in large part due to excessive drive from the subthalamic nucleus. The excessive inhibitory activity of GPi is thought to "brake' the motor thalamus and the cortical motor system to produce the slowness, rigidity, and poverty of movement characteristic of parkinsonian states. To test the hypothesis that direct reduction of Gpi activity can improve motor function, we studied the effect of GPi pallidotomy in 14 patients. The location of the GPi nucleus was confirmed by microelectrode recording before lesion creation. Standardised videotape recordings before and after operation were randomised and scored by a "blinded' evaluator. 6 months after surgery, total motor score in the "off" state had improved by 30% and the total akinesia score by 33%. The gait score in the "off" state improved by 15% and a composite postural instability and gait score by 23%. After surgery there was almost total elimination of drug-induced involuntary movements (dyskinesias), with a 92% reduction on the side contralateral to the pallidotomy. No patient had visual or corticospinal complications. In these patients GPi pallidotomy enhanced motor performance, reduced akinesia, improved gait, and eliminated the neural elements responsible for levodopa-induced dyskinesias.

Helicobacter pylori infection is associated with 95% of duodenal ulcers and more than 80% of gastric ulcers. Several reports have indicated that screening for H pylori may avoid subsequent endoscopic examination. We screened 183 dyspeptic patients, aged under 45, by taking a history of sinister symptoms and regular use of non-steroidal anti-inflammatory drugs (NSAIDs), together with serological testing for H pylori. Endoscopy was performed on 113 patients, of whom 90 (49%) were seropositive, 14 (8%) had sinister symptoms, and 9 (5%) had used NSAIDs regularly. In 34 (19%) patients we detected peptic ulceration. The remaining 70 (38%) patients who were H pylori seronegative, had no sinister symptoms, and had not taken NSAIDs (screen-negative), did not undergo endoscopy but were returned to their primary care physician for treatment of symptoms. At subsequent reassessment (of the non-endoscoped group), symptom severity (p = 0.002), interference with life events (p = 0.01), and medication (p = 0.0002) were all significantly lower in the 6 months after screening than in the 6 month period before screening. Only three screen-negative patients were re-referred after screening but their endoscopic findings were normal. Thus, 67 (36%) endoscopies were avoided. When the non-endoscoped screen-negative patients were compared with a cohort of endoscoped screen-negative patients, the groups did not differ in terms of symptom severity (odds ratio 1.12, 95% CI 0.53-2.35, p = 0.77) or interference with life events (0.82, 0.38-1.76, p-0.61). However, medication use was significantly less (0.37, 0.17-0.80, p = 0.01) in those who did not have an endoscopy. Our study indicates that colonisation screening based on H pylori serology, a history of sinister symptoms, or a history of NSAID use was worthwhile in dyspeptic patients. We avoided 37% of endoscopies and reduced drug usage without disadvantaging those not endoscoped.

Pneumococci are a leading cause of severe bacterial disease in infants and children world wide. A possible means of protecting infants in the first few months of life is immunisation of the mother during pregnancy. We prospectively assessed pneumococcal immunisation of pregnant women to determine the amount of pneumococcal antibody transmitted to the infants in serum and milk and the half-life of the passively acquired antibody. Healthy pregnant women in Dhaka, Bangladesh, were randomised to receive pneumococcal or meningococcal vaccine with routine prenatal tetanus immunisation at 30-34 weeks of gestation. Serum and breast milk specimens from the mothers and sera from infants were collected up to 22 weeks of age and assayed for specific serum IgG, IgG1, and IgG2 and for milk IgA antibodies to pneumococcal serotypes 6B and 19F. 55 mothers and 56 infants were followed from birth to five months. Women who received pneumococcal vaccine had geometric mean antibody increases of 2.6 and 3.4 to types 6B and 19F, respectively. The mean infant/maternal antibody ratios were 0.56 and 0.59 (range 0.11-1.46) for these serotypes. Infant cord antibody titres correlated with maternal titres. Infant/maternal IgG ratios correlated with the interval between immunisation and birth and were higher for specific IgG1 than for IgG2. Infants of pneumococcal vaccine recipients had geometric mean antibody concentrations of 6.8 and 7.5 micrograms/mL to serotypes 6B and 19F in cord blood; in cord blood and in all subsequent serum specimens the concentrations were 2-3 fold higher than in control infants. The median half-life of passive antibody was about 35 days; at five months of age 63-71% of infants of pneumococcal vaccine recipients had antibody concentrations greater than 0.15 micrograms/mL. Breast milk IgA antibodies for pneumococcal serotype 19F, but not for type 6B, were significantly higher in vaccine recipients up to five months after delivery. If maternal pneumococcal polysaccharide antibodies do not interfere with active immunisation of the infant with new glycoprotein conjugate pneumococcal vaccines, passive-active immunisation of infants can be a feasible strategy for developing regions.

Drug resistance has made malaria prevention difficult and the new agents are too expensive for widespread use. Primaquine, an established drug for treatment, is potentially useful for prevention. Malaria prophylaxis with primaquine was evaluated in Irian Jaya during one year in Javanese men who were not deficient in glucose-6-phosphate dehydrogenase (G-6-PD). 126 volunteers were randomised to receive 0.5 mg/kg primaquine base or placebo daily (double-blinded), or 300 mg chloroquine base weekly (open). The protective efficacy of primaquine relative to placebo was 94.5% (95% confidence interval 57-99) for Plasmodium falciparum and 90.4% (95% CI 58-98) for P vivax. Attack rates for either parasite did not differ significantly between the chloroquine and placebo groups. Incidence density of physical complaints not associated with parasitaemia was low (17-18 complaints/person-year) and was about the same in all groups except for cough, which was increased in the primaquine group. Complete blood counts were normal and no evidence of hepatic or renal dysfunction was found with primaquine. However, at 50 weeks the primaquine group had a mean methaemoglobin of 5.8% (range 1.4-13%), which declined by half within 7 days of ending prophylaxis. When used daily for one year by men with normal G-6-PD activity, primaquine was well tolerated and effective for prevention of malaria.

The Coronary Angioplasty versus Bypass Revascularisation Investigation (CABRI) is a multinational, multicentre randomised trial comparing the strategies of revascularisation by CABG (coronary artery bypass grafting) and PTCA (percutaneous transluminal coronary angioplasty) in patients with symptomatic multivessel coronary disease. 1054 patients (820 men and 234 women) were recruited from 26 European cardiac centres. The average age was 60 years and 62% presented with angina of class 3 or greater. 513 patients were randomised to CABG and 541 to PTCA, and 93% and 96%, respectively, of those randomised underwent the allocated procedure. This first report presents data analysed by intention to treat and documents all deaths, major cardiac events, and the symptom status of the patients 1 year after randomisation. After 1 year of follow-up, 14 (2.7%) of those randomised to CABG and 21 (3.9%) of those randomised to PTCA had died. The PTCA group's relative risk (RR) of death was 1.42 (95% CI 0.73-2.76). Patients randomised to PTCA required significantly more reinterventions; only 66.4% reached 1 year with a single revascularisation procedure compared with 93.5% of patients randomised to CABG (RR = 5.23 [3.90-7.03], p < 0.001). The patients in the PTCA group took significantly more medication at 1 year (RR = 1.30 [1.18-1.43], p < 0.001). They were also more likely to have clinically significant angina (RR = 1.54 [1.09-2.16], p = 0.012); this association was present in both sexes but was significant only in females. CABRI is the largest trial of CABG versus PTCA to be reported so far. Its findings are consistent with previous studies, and add to the weight of information that clinicians need to discuss with patients when options for the management of severe angina are under consideration.

We compared different procedures for seeking consent to participate in a sham randomised clinical trial and assessed whether refusal is affected by awareness of the severity of outlook. 2035 healthy subjects aged between 20 and 80 years, who visited a scientific exhibition, were enrolled in a hypothetical trial of experimental versus standard therapy, and randomly assigned to groups asked for conventional informed consent or prerandomisation consent. There were four study groups: one-sided informed consent for randomisation (subjects who refused would receive standard treatment); two-sided informed consent for randomisation (subjects who refused could choose between standard and experimental treatment); randomised consent to experimental treatment (subjects who refused would receive standard treatment); and randomised consent to standard treatment (subjects who refused would receive experimental treatment). The refusal rates were 16.2%, 19.9%, 12.1%, and 49.2%, respectively. The perceived severity of the simulated disease affected the refusal rate: the worse the outlook, the lower the refusal rate for informed consent or for consent after randomisation to new treatment, and the higher the refusal rate for consent after randomisation to standard treatment. The prerandomisation design seems to be efficient in a one-sided clinical scenario (eg, a trial of a new drug that would not be given outside the trial) because the refusal rate was substantially lower for prerandomisation to the new treatment than for conventional one-sided informed consent. However, in a two-sided clinical scenario (eg, a trial comparing similar treatments) the prerandomisation design is potentially highly inefficient; the refusal rate was much higher for prerandomisation to standard treatment than for conventional two-sided informed consent.

In patients with variceal bleeding as a complication of hepatic cirrhosis, propranolol therapy reduces the risk of recurrent variceal haemorrhage. However, the relation between portal pressure response to pharmacological treatment and clinical events has not been well defined. This relation was prospectively investigated in 69 cirrhotic patients receiving continued propranolol therapy after an episode of variceal bleeding. Hepatic venous pressure gradient (HVPG) was measured before and at 3 months of continued drug therapy. At 3 months HVPG had fallen by 20% or more in 25 patients. During follow-up of 28 (SD 17) months rebleeding occurred in 2 of these 25 patients compared with 23 of 44 who had lesser reductions in HVPG. Cumulative probability of rebleeding at 1, 2, and 3 years was 4%, 9%, and 9% in patients with a decrease in HVPG > or = 20%, and 28%, 39%, and 66% in patients with a decrease in HVPG < 20% (p < 0.001, log-rank test). On multivariate analysis, a decrease in HVPG > or = 20% was the only independent predictor of rebleeding (relative risk 0.09, 95% CI 0.02-0.41. Of the 8 patients in whom the HVPG fell to 12 mm Hg or less, none rebled. This study suggests that measurement of the HVPG response to pharmacotherapy will provide useful prognostic information on the long-term risk of variceal rebleeding.

Patients with chronic active hepatitis C and cirrhosis often develop hepatocellular carcinoma. Interferon (IFN) seems to be effective in some patients but whether it prevents carcinogenesis is unknown. In a prospective randomised controlled trial, we evaluated the effects of IFN-alpha in cirrhotic patients with HCV infection because of their high risk of hepatocellular carcinoma. 90 patients with compensated chronic active hepatitis C with cirrhosis were randomly allocated to receive IFN-alpha (6 MU three times weekly for 12-24 weeks) (45 patients) or symptomatic treatment (45 controls), and were followed up for 2-7 years. In nine controls, alanine aminotransferase (ALT) decreased to less than 80 IU/L but did not stay in the normal range. In 19 patients given IFN-alpha, ALT decreased to less than 80 IU/L (in seven patients, it became and stayed normal; p = 0.011, Wilcoxon rank-sum test). However, the mean change in ALT was not significantly different between the two groups. The mean change in peak alpha-fetoprotein values was smaller in patients given IFN-alpha than in controls (p = 0.021). The mean change in the serum albumin level was higher in the IFN-alpha group (p < 0.001). The histological activity index in the 12 IFN-alpha patients undergoing a second biopsy after therapy was improved (p = 0.031). Hepatitis C viral RNA disappeared in seven (16%) of the 45 IFN-alpha patients (95% CI, 7-29%) and in none of the 45 controls (0-8%; p = 0.018). Hepatocellular carcinoma was detected in two (4%, 1-15%) IFN-alpha patients and 17 (38%, 24-54%) controls (p = 0.002, Wilcoxon signed-rank test). The risk ratio of IFN-alpha treatment versus symptomatic treatment was 0.067 (0.009-0.530; p = 0.010 Cox's proportional hazards). IFN-alpha improved liver function in chronic active hepatitis C with cirrhosis, and its use was associated with a decreased incidence of hepatocellular carcinoma.

Joint consultation sessions between general practitioners (GPs) and specialists to examine patients for whom decisions about referral are difficult are thought to be helpful, but their effects have not been evaluated. In a randomised, controlled trial we studied the effects of joint sessions of GPs and orthopaedic surgeons on referral and intervention rates. During 1.5 years, 12 GPs (in groups of three) held monthly joint consultation sessions with four participating orthopaedic surgeons: patients were seen by one orthopaedic surgeon in the presence of three GPs. Patients were included in the trial if the GP was uncertain about the diagnostic or therapeutic management and if referral was considered; and excluded if referral was urgently necessary or if there was some other clear indication for referral. By a randomised consent design, patients were assigned to joint consultation sessions (n = 144) or a usual-care control group (n = 128). A year later the patients were examined by an independent orthopaedic surgeon. There were significantly fewer referrals (51/144 [35%] vs 87/128 [68%], p < 0.01) and diagnostic actions in the intervention group than in the control group, without negative effects on health or functional status. More patients in the intervention group were symptom-free at 1 year (35% vs 24%, p < 0.05). Joint consultation sessions of GPs and orthopaedic surgeons within the framework of general practice resulted in more efficient care, with better targeted examination, treatment, and referrals.

Aspirin is commonly used to treat migraine attacks, although sumatriptan, a much more expensive treatment, is also effective. We compared a combination of lysine acetylsalicylate (equivalent to 900 mg aspirin) and 10 mg metoclopramide (LAS+MTC) with oral sumatriptan (100 mg) and placebo in 421 patients with migraine. LAS+MTC was as effective as sumatriptan with a decrease of headache from severe or moderate to mild or none of 57% and 53%, respectively, for the first migraine attack treated. Both treatments were better than placebo (success rate 24%, p < 0.0001). LAS+MTC was significantly more effective in the treatment of nausea than sumatriptan (p < 0.0001) and was better tolerated (adverse events in 18% and 28%, respectively, p < 0.05). LAS+MTC is as effective as sumatriptan in the treatment of migraine attacks. It is also much cheaper.