eMJA: In Other Journals, 7 April 2008

The declining rate of non-forensic autop-sies has resulted in a loss of important knowledge for clinicians and pathologists, according to Canadian and US researchers. In their review, the authors found that errors in diagnosis which may have resulted in a change in therapy (class 1 errors) are missed in approximately 10% of cases. They also discovered that the expected prevalence of missed cases among non-autopsied patients significantly reduced the rate of antemortem detection of three serious illnesses: aortic dissection, pulmonary embolism, and active tuberculosis. Declining expertise, a lack of resources, and a reluctance to order post-mortem examinations all contribute to the problem. The authors suggest an alternate approach based on the establishment of regional government-funded autopsy centres which would allow proper pathologist training and resuscitate the dying art of the autopsy.

The eyes have it

The relationship between fasting plasma glucose (FPG) and diabetic retinopathy has been questioned and clarified by a group of Australian and international researchers. In a large population-based study involving over 11 000 participants, the prevalence of retinopathy ranged from 9.6% to 15.8%. Retinopathy was diagnosed from multiple clinical photographs, and plasma glucose levels measured via fasting blood samples. Researchers found that there was inconsistent evidence for a glycaemic threshold for retinopathy, with data appearing to suggest a continuous relationship. Of particular concern was the finding that the commonly used FPG cut-off for diabetes of 7.0 mmol/L had a sensitivity of less than 40% for the detection of retinopathy.

Lancet 2008; 371: 736-743

Down time for teens

The results of a US randomised controlled trial suggest that adolescents with selective serotonin reuptake inhibitor (SSRI)-resistant depression may benefit from a change to another antidepressant combined with cognitive behaviour therapy (CBT). A group of 334 teenagers with a diagnosis of major depressive disorder who had not responded to a 2-month trial of an SSRI were enrolled in the trial. Participants were randomly assigned to four groups receiving either: a different SSRI, a different SSRI plus CBT, a change to venlafaxine, or venlafaxine plus CBT. Therapy was continued for 12 weeks. CBT plus a switch to either medication showed the highest response rate, but those on venlafaxine had a greater incidence of side effects including an increase in diastolic blood pressure and the occurrence of skin problems.

JAMA 2008; 299: 901-913

Depressed doctors — who pays?

Hospital residents working in paediatrics have high rates of depression and “burnout” which may affect their prescribing ability, US researchers have found. In a prospective cohort study involving 123 residents in three paediatric residency programs, the prevalence of depression and burnout was assessed using validated rating scales. The rate of medication prescribing errors per month for each resident was also determined. In total, 20% of the participating doctors met the criteria for depression and 74% for burnout. Depressed residents were found to make over 6 times as many medication errors as non-depressed residents. The rates of errors per month did not differ for the residents suffering burnout. The authors call for screening of residents for signs and symptoms of depression, and warn that the alarming results of their study indicate a need to improve the mental health of trainees in the medical field.

BMJ Online, 7 Feb 2008

Variable LFTs create confusion

Liver function tests can show considerable variability in the same asymptomatic individual over time, resulting in unnecessary further investigations and patient anxiety, according to the results of a large population study. The US NHANES (National Health and Nutrition Examination Survey) included over 1800 participants in a reliability study aimed at characterising the magnitude of intra-individual variability of liver function test results. Tests included levels of bilirubin, aspartate aminotransferase (AST), alanine aminotransferase, alkaline phosphatase and γ-glutamyltransferase. A significant percentage of results that were initially elevated were normal on a second examination (a mean of 17.5 days later), including 38% for bilirubin and 36% for AST. The analysis took into account possible confounders such as alcohol consumption, hepatitis, or use of medications known to be hepatotoxic. The authors suggest that clinicians need to be aware of the high variability of these tests and consider retesting asymptomatic individuals before proceeding to other, more expensive investigations.

Ann Intern Med 2008; 148: 348-352

Dr Tanya Grassi, MJA

The Medical Journal of Australia eMJA