BMJ  2006;332:1264-1265 (27 May), doi:10.1136/bmj.332.7552.1264

Practice

Short cuts

What's new in the other general journals

Meta-analysis finds rare but serious side effects

Randomised trials are rarely big enough, long enough, or powerful enough to give patients the information they want about rare but serious side effects. In an unusual move, researchers used meta-analysis, a powerful tool for combining trial results, to find out more about the cancer causing potential of a treatment for rheumatoid arthritis—the monoclonal antibodies infliximab and adalimumab. Both are directed against tumour necrosis factor.

View larger version (55K): [in this window] [in a new window]   Credit: JAMA

Pooled data from nine placebo controlled trials, which included more than 5000 patients, showed that the antibodies were associated with a significantly greater risk of cancer (odds ratio 3.3, 95% CI 1.2 to 9.1) and of serious infections (2.0, 1.3 to 3.1). Overall, 24 of the 3493 patients given monoclonal antibodies developed cancer (0.8%), compared with three of the 1512 given placebo (0.2%). Higher doses were associated with higher risk of cancer.

These findings are consistent with what we already know about tumour necrosis factor—it helps combat infections and kills tumour cells—and with what researchers suspected might happen when it was disabled. The results are also consistent with previous hints from postmarketing surveillance, cohort studies, and individual trials.

The authors say they are among the first to use meta-analysis to assess a drug's safety, and it works well. They suggest others follow suit.

JAMA 2006;295: 2275-85[Abstract/Full Text]

Low oxygen and pressure don't promote clotting

Fifty years after someone first noticed the connection between air travel and venous thrombosis, scientists are still trying to untangle the complex relationship between haemostasis, sitting still for a long time, and the combined effects of low pressure and low oxygen tension in the cabin. In the latest controlled experiment, hypobaric and hypoxic conditions mimicking a long haul flight did not alter volunteers' haemostasis any more than sitting for eight hours in conditions mimicking a deck chair in the garden at 50-70 m above sea level. Serum markers for coagulation, fibrinolysis, platelet function, and endothelial activation changed a little during both halves of the experiment, in part because of documented diurnal variations in haemostasis.

View larger version (60K): [in this window] [in a new window]   Credit: ANN INT MED

So if sitting still for long periods in low air pressure with a reduced arterial oxygen saturation is not thrombogenic, what is? An editorial (pp 2297-9) thinks it's likely to be some kind of interaction between the conditions in an aeroplane's cabin and other risk factors for venous thromboembolism. People with a genetic predisposition to clotting were excluded from this experiment, but other experiments in people at high risk show just such interaction between mild hypoxia, use of oral contraceptives, and factor V Leiden mutations.

JAMA 2006;295: 2251-61[Abstract/Full Text]

Low pressure headaches are underdiagnosed and undertreated

Doctors don't know much about spontaneous leaks of cerebrospinal fluid, which is why they often misdiagnose the ensuing headache as, migraine, viral meningitis, tension headache, or even malingering, says a review. It can take years to get it right.

The headache, which classically gets worse when the patient stands up, is caused by low pressure cerebrospinal fluid around the brain. The brain sinks down, pulling on pain sensitive structures, such as the dura. The true incidence is unknown, but in emergency departments, low pressure headaches are about half as common as spontaneous subarachnoid haemorrhage. Women seem more likely to develop the leaks than men, and the peak age is about 40 years.

The associated headaches are frequent and often debilitating. Look out for the patient carrying a pillow into the surgery so they can lie down in comfort. Other diagnostic clues include neck pain, nausea, and vomiting, which the author says are reported by about half of patients with low pressure headaches. They can also have altered hearing, get tinnitus, or be "off balance." Myelography, which exposes the leak, is the most accurate test, but cranial magnetic resonance imaging is less invasive.

There are plenty of treatments, including bed rest, epidural blood patches, fibrin sealant, and surgery to repair the leaks. None have been tested in clinical trials.

JAMA 2006;295: 2286-96[Abstract/Full Text]

Simple questions help GPs diagnose incontinence

About a third of women of at least 40 years old have some kind of urinary incontinence. Most of it goes untreated. To encourage primary care doctors to look for and treat incontinence, US researchers developed a quick and easy questionnaire for them to use. It takes about 30 seconds and asks "during the last three months, have you leaked urine (even a small amount)?" followed by two multiple choice questions characterising the incontinence as urge incontinence, stress incontinence, other, or mixed.

Compared with the extensive evaluation recommended by US and other guidelines, their simple questionnaire was fairly accurate. It diagnosed stress incontinence with a sensitivity of 0.86 (95% CI 0.79 to 0.90) and a specificity of 0.60 (0.51 to 0.68) and diagnosed urge incontinence with a sensitivity of 0.75 (0.68 to 0.81) and a specificity of 0.77 (0.69 to 0.84). The distinction is important, because the treatment is different for different types.

The researchers tested their questionnaire in an ethnically mixed and unselected cohort of US women with previously undiagnosed incontinence. The authors say it's not perfect but should be good enough to guide initial treatment in primary care pelvic floor exercises for stress incontinence or urge suppression exercises for urge incontinence. If simple treatments don't work they suggest referring women for more sophisticated and invasive investigations after six to 12 months.

Ann Int Med 2006;144: 715-23[Abstract/Full Text]

Detect hearing loss early for better language skills later on

Screening newborn babies for impaired hearing has measurable effects on their language later in childhood. In an observational study of 120 children with hearing impairment, children screened at birth were diagnosed as having hearing loss earlier than children who were not screened. Being diagnosed before the age of 9 months was associated with better receptive and expressive language skill at about 8 years, and was also associated with a better match between verbal and non-verbal ability. The simple act of screening was associated with better receptive language skills. Screening and early detection did not improve children's speech, however, which was measured by questioning the parents. The authors say this was probably an insensitive way of assessing speech and are planning a later study using tape recordings of the children instead.

View larger version (41K): [in this window] [in a new window]   Credit: N ENGL J MED

The benefits of confirming hearing loss before 9 months looked clinically meaningful in this study—equivalent to an increase of about 10 or 12 points in verbal IQ. The advantage may be even greater now that children have less time to wait for their hearing aids. The children in the study were among the first to be screened in the United Kingdom, when the infrastructure for rapid confirmation and treatment was still evolving.

N Engl J Med 2006;354: 2131-41[Abstract/Full Text]

Robots may be the future

First there was surgery, then came keyhole surgery, and now robots are poised to take surgeons into even smaller spaces through even tinier holes, writes one enthusiast. More than 10 000 operations have already been done by remotely operated robots (called da Vinci, ZEUS, AESOP, or Robodoc), and surgeons, particularly paediatric surgeons, dream of the day they can get wherever they want to go from a wrap around console a safe distance—or even any distance—from the blood and guts.

Surgical robots are not cybermen, he says. Most are a collection of tools interfaced digitally with a computer. The tools are more sophisticated than laparoscopic instruments, which can go only backwards, forwards, clockwise, or anticlockwise. Robotic tools have wrists. They tilt, grip, and move up and down, guided by a surgeon with a joystick. So far, surgeons have removed prostates, replumbed coronary arteries, and done a tricky bile duct anastomosis ina6weekold baby, all by robot.

New technologies are not always better than old ones, however, and, as usual, proper evaluation is lagging far behind. Robots are expensive, and it takes a long time to learn to use one safely. It will take more than a few case studies and some technical notes to convince health providers to invest.

N Engl J Med 2006;354: 2099-101[Full Text]

Prevalence of hepatitis C stays high in the US

Millions of people in the United States remain chronically infected with hepatitis C virus, write researchers, who estimate that 1.6% of the adult population are infected. Their study shows that a history of injecting drugs is by far the most powerful risk factor, with an odds ratio of nearly 150 (95% CI 45 to 494). Promiscuous sex and a history of blood transfusion before 1992 (for people aged at least 60 years) were also associated with infection.

These data came from the latest national health and nutrition examination survey. Between 1999 and 2002, more than 15 000 US adults gave a detailed medical history as well as blood for serology and other tests.

View larger version (44K): [in this window] [in a new window]   Credit: ANN INT MED

An editorial (pp 770-1) says the prevalence of hepatitis C infection has remained fairly constant since the last survey, more than 10 years ago, despite a dramatic reduction in the incidence of new infections. The current trends were set in the 1960s and '70s, when a cohort of "baby boomers" was infected, possibly while experimenting with drugs. This cohort is gradually ageing, says the editorial, and will start to die off about 2015. Prevalence will not start to decline until then. In the meantime, doctors should expect a "bitter harvest" of chronic liver disease.

Ann Int Med 2006;144: 705-14[Abstract/Full Text]

Two antiplatelet agents work better than one after stroke

Aspirin helps prevent further vascular events in patients with a history of stroke. The combination of aspirin and dipyridamole works even better, a large trial has found. Compared with aspirin alone, the two treatments together reduced the combined risk of death from a vascular event, non-fatal heart attack, stroke, or serious bleeding by three percentage points from 16% to 13% over 3.5 years (hazard ratio 0.80, 95% CI 0.66 to 0.98). That's a number needed to treat of 104.

This large trial is the sixth to try and find out whether adding dypiridamole to aspirin is a safe and effective preventive strategy for people with vulnerable cerebral arteries. The authors helpfully added their results to their existing meta-analysis of the previous five to give a final risk ratio of 0.82 (0.74 to 0.91) for people taking both. An editorial (pp 1638-9) says that these findings are about as clear as they can be, and adds that combined treatment should now be an option for people with a history of ischaemic stroke or transient ischaemic attack. Doctors can be reassured that the two antiplatelet agents (about 75 mg a day of aspirin plus 200 mg twice a day of dipyridamole) caused no more serious bleeding than aspirin alone, although other side effects, particularly headaches, were more of a problem for people taking both.

Lancet 2006;367: 1665-73[CrossRef][ISI][Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?