Short Cuts

What’s new in the other general journals

BMJ 2007; 335 doi: (Published 13 December 2007) Cite this as: BMJ 2007;335:1234

Warfarin, insulin, and digoxin cause a third of all serious side effects in over 65s

Hospital emergency departments in the US have to deal with more than 177 000 visits a year from older people with an adverse drug event. In an analysis of routine surveillance data, oral anticoagulants or antiplatelet agents, antidiabetic drugs, and drugs with a narrow therapeutic index accounted for nearly half of all visits (47.5%, 95% CI 40.2% to 54.8%). Just three drugs—warfarin, insulin, and digoxin—accounted for a third of visits (33.3%, 27.8% to 38.7%).

The risk associated with these three commonly prescribed drugs was around 35 times higher than that associated with the official list of 41 other drugs known to be potentially dangerous for older people. This list, known as the Beers criteria, is widely used to gauge the quality of prescribing. Quality indicators that focus on warfarin, insulin, and digoxin may be more useful, say the researchers. Small improvements in prescribing and monitoring of these drugs could have an important effect on the incidence of adverse drug events in people over 65.

All the events in this analysis were serious enough to justify visiting an emergency department. Clinically obvious bleeding was the most common adverse event associated with warfarin (73% of events). Hypoglycaemia, often with unconsciousness or seizures, was the most common event associated with insulin (95.4%).

Antibiotics and intranasal steroids don’t help in sinusitis

Patients presenting to their primary care doctor with symptoms typical of bacterial sinusitis are unlikely to get any relief from antibiotics or topical corticosteroids according to a randomised trial from the UK. Neither treatment—alone or in combination—improved symptoms better or faster than placebo.

The 240 adults took amoxicillin 500 mg three times daily for seven days, 200 μg a day of budesonide into each nostril for 10 days, both treatments, or a double placebo. Just over two thirds of participants were completely better within 10 days, whatever their treatment.

The inclusion criteria for this trial were selective and based on strict criteria, such as purulent nasal discharge and pain, which had previously been validated for use in secondary (but not primary) care. The authors say antibiotics and nasal steroids are even less likely to work in real world patients with sinusitis, many of whom have viral infections. Recruitment was slow, but the authors are confident that they randomised enough patients to exclude any clinically important benefits.

An editorial (p 2543) says their findings are consistent with previous work and urges doctors to exercise caution when considering active treatments for patients with possible acute sinusitis. It won’t be easy. More than 90% of such patients in the UK and US currently receive antibiotics.

Faith in vitamin E, β carotene, and oestrogen endures despite the evidence

Throughout the 1990s many scientists, doctors, and patients believed that vitamin E could help prevent cardiovascular disease. Large observational studies found clear evidence of benefit. Since then, equally large and more convincing randomised trials have found the opposite. Belief in the protective properties of vitamin E persists, however. A close look at scientific articles citing the early observational studies showed that half were still arguing in favour of vitamin E as late as 2005, five years after a landmark trial overturned the observational studies’ findings.

Belief in the anti-cancer potential of β carotene has endured even longer. The notion that β carotene prevents cancer was discredited by randomised trials published in 1996. In another citation analysis, two thirds of citations to the original observational work were still favourable 10 years later. Most of the articles simply ignored the later evidence. Results were similar for the more recently discredited idea that oestrogen protects postmenopausal women against dementia.

In the vitamin E analysis, articles in specialist journals were more likely than articles in general journals to defend the protective properties of vitamin E. So were articles that reported their own observational data, those from outside the US, and those that failed to cite the key (negative) randomised trial.

Fat children are more likely to get heart disease as adults

The efficiency of Danish registers at recording people, deaths, and hospital discharges over long periods of time has allowed researchers to confirm and fine tune previously suspected associations between childhood weight and coronary heart disease later in life. In a cohort study of more than 276 000 schoolchildren, risk of heart disease in adulthood rose in line with adjusted body mass index (BMI) at ages 7 to 13. For example, each extra unit of BMI z score (equivalent to a weight increase of 6.3 kg) in a 13 year old boy was associated with a 24% increase in the risk of fatal heart disease in adulthood and a 17% increase in the risk of a non-fatal event (hazard ratios 1.24, 95% CI 1.19 to 1.29 and 1.17, 1.14 to 1.20).

The risks associated with being overweight in childhood went up with age but were detectable and significant even at 7 years for boys and 9 years for girls. A higher BMI z score was systematically riskier for boys than for girls at all ages.

Children and adolescents continue to get heavier in Denmark and elsewhere, say the researchers, and the trend is likely to continue for the foreseeable future. If it does, a companion study modelling the epidemic in the US estimates that by 2035 the prevalence of coronary heart disease there will have increased by 5-16% (p 2371).

Risk of death from pulmonary embolism low after anticoagulation for venous thromboembolism

Adults who complete their anticoagulant treatment after a first episode of venous thromboembolism have a low risk of dying from pulmonary embolism, according to a cohort study that tracked patients for a mean of four and a half years.

The overall risk of a fatal pulmonary embolism was well under 1% a year and lay somewhere between 0.49 events (95% CI 0.36 to 0.64) per 100 person years and 0.19 events (0.12 to 0.30) per 100 person years, depending on the certainty of the cause of death. Risks were similar for patients who had been treated for deep venous thrombosis or for pulmonary embolism. The same study found that for patients with a recurrence, the risk of dying from it was 4-9%, again depending on the certainty of the cause of death.

The 2052 participants in this cohort were from two previous studies. Most (96%) had completed three to 12 months of treatment, usually for an isolated deep vein thrombosis (1450/2052, 71%). Just under half of participants had an underlying risk factor such as lower limb trauma or recent surgery. Those with active cancer, permanent immobility, or high risk thrombophilia were excluded from both source studies.

Baclofen helps people with alcoholic liver disease stop drinking

Baclofen could be a much needed new treatment for people with alcoholic liver disease. A small placebo controlled trial suggests it can reduce alcohol cravings and help people stop drinking without doing any further damage to the liver. Over 70% of the 42 Italian adults who took baclofen for 12 weeks managed to stop drinking completely, compared with only 29% of the 42 who took placebo (odds ratio 6.3, 95% CI 2.4 to 16.1). The baclofen group had significantly more drink free days (63 v 31, P=0.001), significantly lower scores on a validated measure of alcohol craving, and significantly fewer relapses. Their liver function tests improved during the study. Baclofen is excreted by the kidneys, unlike other drug treatments for alcohol dependence such as naltrexone, say the authors. So it should be safer for people with advanced liver disease. All the participants in this trial had cirrhosis and were still drinking when recruited from a single Italian clinic. They all had counselling in addition to their allocated treatment.

An accompanying comment welcomes the results, which are promising enough to justify bigger, longer, and more diverse trials of baclofen (p 1884). It is particularly important to find out whether baclofen might work for the many people with alcoholic liver disease who have mental illness and who misuse other drugs, says the author.

Treatment failure develops slowly in UK adults with HIV

Adults in the UK who start treatment for HIV with at least three antiretroviral drugs have a low risk of extensive treatment failure, say researchers from London. They recently estimated that the three main classes of drug will continue to work for at least 10 years in more than 90% of patients. Their cohort study of nearly 8000 adults found that the cumulative risk of all three drug classes failing was only 9.2% (95% CI 5.0% to 13.4%) 10 years after the start of treatment.

The researchers defined failure as a viral load of more than 400 copies/ml despite more than four months of treatment. By extensive failure they meant that all three subclasses of nucleoside reverse transcriptase inhibitor, at least one non-nucleoside reverse transcriptase inhibitor, and a ritanovir boosted protease inhibitor all failed.

While these results are good news for adults in the UK and other developed countries, they are unlikely to translate well to poorer countries where most people with HIV live, says a linked comment (p 1885). For example, in Africa people have more advanced disease before they start treatment and drug choices are limited. When first line treatments fail fewer options are left. Research on treatment failure in these settings is badly needed. So is more equitable access to the alternatives enjoyed by the wealthy few.

View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to to receive unlimited access to all content on for 14 days.
Sign up for a free trial