Short Cuts

All you need to read in the other general journals

BMJ 2008; 336 doi: (Published 06 March 2008) Cite this as: BMJ 2008;336:530

Resurrect autopsies for the benefit of the living

Autopsies are dying out in the US. Fewer and fewer are done each year—even in the big teaching hospitals—and if the trend continues to its natural conclusion, doctors, patients, and the wider public will lose an important source of information about how and why people die, say two observers. Autopsies fine tune diagnoses and sometimes change them altogether. They help doctors improve their diagnostic skills and provide hospitals with an important tool for improving the quality and delivery of diagnostic imaging and laboratory services. It is also important for relatives to know that their father or brother died from an aortic dissection not a heart attack, one of the most common mistakes reported by autopsy studies.

Pathologists are losing interest in autopsies, new trainees are failing to learn the techniques, and quality is suffering. The lack of financial incentives for hospitals to perform autopsies hasn’t helped.

These experts say it is too late and too uneconomical to reverse the trend completely, but it may be possible to concentrate the necessary skills in regional centres. These hospitals could help maintain a service that benefits the living by improving the accuracy of death certificates, increasing the reliability of studies analysing causes of death, and teaching doctors where they have gone wrong.

Erythropoietins linked to excess deaths in people with cancer

Further evidence has emerged that synthetic erythropoietins can reduce survival in patients with anaemia and cancer. A meta-analysis of 51 phase III trials reports an overall hazard ratio for death of 1.10 (95% CI 1.01 to 1.20) for patients given epoetin or darbepoetin compared with placebo. The same agents were associated with a larger 57% increase in the risk of venous thromboembolism (relative risk 1.57, 1.31 to 1.87) in an analysis of 38 trials.

The US Food and Drug Administration has issued a series of warnings about the potential hazards of epoetin and darbepoetin in patients with both cancer and chronic renal failure. The regulator is scheduled to take another close look at the evidence in a few days. Oncologists regularly prescribe these agents to patients with anaemia secondary to cancer treatments such as chemotherapy. They have also been tested in patients with cancer related anaemia who are not receiving chemotherapy. An adverse effect on mortality is biologically plausible, say the authors, although the precise mechanism is unclear. We know that some human cancers express erythropoietin receptors. Stimulating them activates signalling pathways that cause cell proliferation, reduced apoptosis, and invasion.

Doctors should be environmentalists too

Climate change is likely to be bad for human health. Drought, mass migration, food shortages, a surge in vector borne diseases, and deaths from extreme weather are just some of the possibilities, writes one surgeon from the US. Doctors and their institutions must do more to educate themselves about these very real threats and help to educate others. Medical schools could teach environmental science, specialist societies and research institutes could gather evidence on the likely effects of accelerating climate change on human health, hospitals could go green, and doctors could join forces with environmentalists to make plans and take action, he says.

The medical profession has a long history of responding to global challenges such as epidemics and natural disasters. It is time to step up and confront the state of the environment, along with the world’s governments, scientists, and businesses.

No one knows exactly what will happen, or how fast. But we do know that the ice caps are melting, the ozone layer is thinning, the planet is warming, and the world’s forests are being deliberately torched to grow crops to feed an exponentially increasing human population. All these events damage human health in one way or another, he says. Doctors must be part of the response.

Complex interventions help older people remain at home

Systems of care for older people tend to be driven by cost containment rather than evidence. So researchers from the UK did a detailed meta-analysis to find out if multifactorial interventions do what they are supposed to do, and help older people live safe and independent lives in the community.

Overall, the results were positive. The complex interventions they looked at reduced nursing home admissions (relative risk 0.87, 95% CI 0.83 to 0.90), hospital admissions (0.94, 0.91 to 0.97), and falls (0.90, 0.86 to 0.95). The effects were small, but likely to be worthwhile in the real world, says a linked comment (p 699). Multifactorial interventions had no effect on deaths.

All the interventions included some kind of assessment and home visits but were otherwise very varied. This analysis was unable to pin down the most effective aspects of assessment and care. The number of professionals involved and the intensity of home visits seemed to make no difference to overall outcome. We shouldn’t be surprised given the variety of older people included in these 89 trials, says the comment. Different people have very different vulnerabilities. Care should always be tailored closely to need.

A new glucose threshold for diabetes? Not yet

The World Health Organization and the American Diabetes Association define diabetes as a fasting plasma glucose concentration of 7 mmol/l or higher. They chose this threshold because epidemiological studies suggested it could differentiate people who are at risk of microvascular complications from people who weren’t. A new study looking exclusively at retinopathy found no evidence of a threshold effect, and it concluded that the association between glucose concentration and retinopathy is more likely to be linear and to continue below 7 mmol/l. The authors analysed data from three contemporary cross sectional studies. Retinal photographs showed signs of retinopathy in up to 13.4% of participants without diabetes. They say opinion leaders should look again at the accepted diagnostic threshold.

An accompanying commentary isn’t so sure (p 700). We know that hyperglycaemia is harmful with no definable lower threshold for risk of cardiovascular or renal disease, it says. The currently accepted cut-offs for glucose concentration may be flawed, but they are the best we can do given the serious limitations of the data available. Despite these latest findings, the 7 mmol/l cut-off probably does define a group of people with an increased risk of harm from hyperglycaemia. We should stick with it until a better diagnostic tool comes along.

Switching drugs and adding CBT improves response of teenagers with refractory depression

When depressed adolescents don’t respond to first line treatment with a selective serotonin reuptake inhibitor (SSRI), adding cognitive behaviour therapy (CBT) and switching to another drug gives them the best chance of improvement, a randomised trial has found.

The 334 teenage participants had moderately severe enduring depression despite an average of 17 weeks’ treatment with an SSRI and sometimes psychotherapy other than CBT. More than half had suicidal ideation. The chance of an adequate response over 12 weeks was significantly higher in participants who started CBT and switched either to venlafaxine or to another SSRI, compared with those who just switched drugs (response rates 55.8% (91/166) v 40.5% (68/168), P=0.09). The two drug options worked equally well in combination with CBT, but venlafaxine caused more side effects—including increased diastolic blood pressure, a higher pulse rate, and more skin problems—than a new SSRI.

The authors combined two standard measures to define an adequate clinical response—a fall of at least 50% in symptom scores obtained by interview and a global impression of improvement. Choice of treatment had no effect on suicidal ideation or self rated symptoms, both of which improved over time.

Vasopressin fails to reduce mortality in patients with septic shock

Vasopressin is a small peptide hormone that acts as a vasoconstrictor and may have a useful role in people with septic shock. A low dose had no effect on mortality in the latest large trial, however. The researchers compared a low dose of vasopressin added to noradrenaline with noradrenaline alone. Patients were also allowed other open label vasopressors if required. Mortality at 28 days was 35.4% (140/396) in the vasopressin group and 39.3% (150/382) in controls (P=0.26). Ninety day mortality was also unaffected by vasopressin.

The authors were disappointed. They were also puzzled when a subgroup analysis suggested that vasopressin reduced mortality in patients with the least serious disease. They were expecting the opposite and say these results should be treated with caution.

Vasopressin is already widely used and in this trial did no discernible harm. It did no discernible good either, perhaps because the patients were treated too late, says an editorial (p 954). Evidence is mounting that the timing of treatment is as important for patients with shock as it is for patients with heart attack—the sooner the better. These patients began treatment a mean of 12 hours after the onset of septic shock.

Are we overusing antibiotics in people with severe dementia?

Nursing home residents with severe dementia are prescribed a lot of antibiotics, particularly as they approach death, say researchers. Of the 214 residents in their cohort, 142 (66.4%) had one or more courses of antibiotics during 18 months of follow-up. Residents who were treated had an average of four courses each. Quinolones and third generation cephalosporins were used most often, and the most common indication was chest infection (252/540, 46.7% of courses), followed by urinary tract infection (192/540, 35.6%). Almost half the participants died during the study. Antibiotic use went up significantly in the final few weeks before death.

The researchers and a linked editorial (p 349) agree that such extensive use of broad spectrum antibiotics in this population is worrying. There is no evidence that treatment prolongs life or relieves distress in people with severe disabling cognitive impairment. It does, however, increase the spread of antimicrobial resistance. This small observational study cannot tell us whether antibiotics are being overused in people with end stage dementia, says the editorial. But we should at least be asking the question.