Short Cuts

All you need to read in the other general journals

BMJ 2009; 338 doi: (Published 16 June 2009) Cite this as: BMJ 2009;338:b2417

Uncertainty continues over steroids for sepsis

Corticosteroids have a long history in the treatment of septic shock. The evidence has gone first one way, then the other, with dedicated critical care doctors not far behind, says an editorial (pp 2388-90). In 2002, the United States briefly ran out of hydrocortisone after an unusually rapid and enthusiastic response to a positive French trial of low dose therapy. More recently, longer courses of low doses have fallen out of favour. Current guidelines sit on the fence with “suggestions” rather than recommendations, citing weaknesses in the evidence for their uncertainty.

The latest addition to this messy situation is a meta-analysis of all potentially relevant trials, including 12 trials that looked specifically at the benefits of at least five days of low dose corticosteroids. This low dose strategy was associated with lower mortality at 28 days than a high dose approach (236/629 (37.5%) v 264/599 (44.1%); risk ratio 0.84, 95% confidence interval 0.72 to 0.97).

The authors of this study and the authors of the editorial agree this subanalysis cannot be the final word, however. Other analyses of all trials (short and long duration of treatment; high and low doses) failed to find conclusive evidence of benefit.

Clearly, health professionals caring for patients with severe sepsis and septic shock will have to live with uncertainty for at least a little while longer, says the editorial. Doctors should embrace this uncertainty, then share it with patients and their relatives. For now, “the final decision rests squarely on those at the bedside.”

Six months of androgen suppression is not enough for men with locally advanced prostate cancer

Long term androgen suppression prolongs survival in men who have had radiotherapy for locally advanced prostate cancer. Treatment with analogues of luteinising hormone releasing hormone causes unpleasant side effects though, including poor sex drive, hot flushes, heart attacks, and fractures. Can men safely stop androgen suppression after just six months?

A trial comparing short term with long term androgen suppression reported a clear result in favour of continuing treatment for at least three years. Men who stopped treatment after six months had significantly higher all cause mortality over five years than men treated for longer (19% v 15.2%; hazard ratio 1.42, 96% CI 1.09 to 1.85). Short term androgen suppression was also associated with more deaths from prostate cancer (1.71, 95% CI 1.14 to 2.57). The trial was designed to test the “non inferiority” of short term treatment, but the treatment failed the test at an interim analysis and the researchers stopped recruiting early.

Most of the 970 participants had clinical stage T2c or T3 cancer. Radiotherapy followed by long term androgen suppression looks like a good option for these people, despite the side effects, says an editorial (pp 2572-4). It’s less clear how we should treat the growing number of men who have smaller and less advanced cancers detected following screening tests for prostate specific antigen.

Reassuring data for women taking metoclopramide in pregnancy

Metoclopramide is an effective antiemetic for women in early pregnancy and doctors in some countries use it routinely. Safety isn’t yet established, but a data linkage study from Israel looks reassuring. The authors analysed a cohort of 81 703 women who gave birth at one centre over a 10 year period. All were members of Israel’s largest health maintenance organisation and about two thirds were Bedouin.

Just over 4% (3458/81 703) of the cohort took metoclopramide during the first trimester of pregnancy. These women were no more likely to have babies with major congenital malformations than women who had not taken metoclopramide in the first trimester (264/599 (5.3%) v 3834/78 245 (4.9%); odds ratio 1.04, 95% CI 0.89 to 1.21). The drug was not associated with minor malformations (1.10, 0.92 to 1.31), low birth weight (1.01, 0.89 to 1.14), preterm delivery (1.15, 0.99 to1.34), perinatal death (0.87, 0.55 to 1.38), or low Apgar scores (0.84, 0.57 to 1.22) after adjustment for half a dozen potential confounders including smoking, age, and parity. The authors could not find any evidence of link between higher doses taken in early pregnancy and higher risk. They had no data on spontaneous abortions.

Best practice reduces radiation dose from cardiac computed tomography

In September 2007, 15 hospitals and imaging providers in Michigan, US, implemented a strategy to minimise the radiation dose associated with cardiac computed tomography angiography. Elements included minimising the range of each scan, synchronising the tube current with the cardiac cycle (gated current modulation), using β blockers to control heart rate and variability, and reducing the scan voltage where possible.

In this prospective before and after study of 4862 patients, average radiation dose per patient fell by more than half from 21 mSv to 10 mSv (P<0.001). Study sites had eight months to implement the programme, including training and educating staff. In the two months after full implementation, the most common dose received by patients was 5-10 mSv, compared with 25-29 mSv in the control period. Reducing the scan voltage emerged as the most powerful factor associated with achieving target radiation doses less than 15 mSv. The dose reduction strategy made no discernible difference to the quality of the scans, which were assessed by doctors at study sites. Around 90% of all scans—both before and after the dose reduction programme—were good enough to be diagnostic.

Cardiac computed tomography is a useful diagnostic tool for heart disease, say the authors, who believe radiation exposure can be controlled with best practice. They concede that other providers may find it harder to implement changes than their dedicated and enthusiastic consortium.

Urgent pages often go to the wrong doctor

Communication failures can be a serious threat to patient safety, so researchers decided to explore the efficiency of the paging systems at two teaching hospitals in Toronto. They reviewed all 10 190 pages sent to 28 residents in both hospitals during a two month period in 2008. One in seven pages (1409/10 190 (14%)) had been sent to the wrong doctor—defined as a resident who was scheduled to be off duty, out of the hospital, on study leave, or on holiday.

A closer look at 213 of these misdirected pages showed that nearly half (47%) were at least urgent and 15% reported clinical emergencies such as a low oxygen saturation or a high serum potassium concentration. One simply said “call radiology-alert perforation.” Most of the mistakes were made while paged residents were resting after a period on call, taking the evening off, or scheduled for an academic half day.

This quick retrospective analysis can’t tell us much about what happened to the patients, but it does flag up inefficiencies that must have caused delays and disrupted work flow, say the researchers. They are “taking steps” to rectify the problem.

Progesterone doesn’t prevent early delivery of twins

It’s now fairly clear that progestogens do not help prevent preterm delivery of twins, say researchers. In their placebo controlled trial, a vaginal gel containing progesterone had no effect on the risk of delivery or intrauterine death before 34 weeks’ gestation (61/247 (24.7%) v 48/247 (19.4%); odds ratio 1.36, 95% CI 0.89 to 2.09). Women used an active or placebo gel daily from week 24 of gestation. A combined analysis of two other published trials testing progestogens in twin pregnancies produced the same negative result.

Progestogens may work better in singleton pregnancies, say the authors, although it’s still unclear whether delaying delivery improves long term outcomes for babies. One commentator is particularly concerned about this hole in the evidence and urges researchers, sponsors, and grant giving bodies to recognise that treatments to prolong pregnancy could harm babies as well as help them (doi:10.1016/S0140-6736(09)61077-1). Trials must, therefore, include long term follow-up to assess health and wellbeing beyond delivery and well into childhood. Trial registers list another 17 trials in the pipeline evaluating progestogens to prolong pregnancy in either singleton or multiple pregnancies, he writes. Only two plan any kind of long term follow-up, and all are too small to assess the balance of risks and benefits with any certainty.

Revascularisation is unnecessary for many patients with diabetes and heart disease

The latest trial investigating the best treatment strategy for people with type 2 diabetes and stable heart disease used a factorial design to compare revascularisation with best medical treatment for heart disease while at the same time comparing insulin sensitisation with insulin provision for diabetes. Overall, choice of treatment for either heart disease or glucose control made no significant difference to risk of death or cardiovascular events over 5 years in 2368 adults.

Important differences emerged from secondary analyses, however. Early revascularisation with coronary artery bypass surgery prevented a significantly greater number of serious cardiovascular events, particularly heart attacks, than best medical treatment alone (22.4% v 30.5%; P=0.01), whereas early revascularisation with stenting did not. Patients suitable for coronary artery bypass surgery had worse heart disease than patients suitable for percutaneous coronary intervention, and the two subgroups were randomised separately. Among the 763 patients suitable for coronary artery bypass surgery, those who had both surgery and treatment with insulin sensitising agents such as metformin and rosiglitazone had the lowest risk of serious cardiovascular events over five years (18.7%).

This study and other trials suggest that, in general, medical therapy rather than invasive revascularisation is a good first line option for many patients with diabetes and moderate, stable heart disease, says an editorial (pp 2570-2). Best medical treatment for these patients included β blockers, statins, angiotensin converting enzyme inhibitors or angiotensin receptor blockers, and aspirin. When revascularisation is unavoidable, coronary artery bypass surgery should be the preferred strategy.

The global research effort is inefficient, wasteful, and ignores consumers

Globally, biomedical research costs billions of US dollars each year, and much of it is wasted, say two commentators. Researchers ask the wrong questions using the wrong methods, results are poorly reported and selectively published, and research funding is misdirected, ignoring the priorities of patients who want treatment options beyond drugs. Much of the output is still inaccessible to the public, and duplication of effort continues to be a problem. Over half of clinical trials are designed without reference to previous research on the same question. Taken together, inefficiencies across all stages of the process probably waste between 50% and 85% of global research funds, they write.

Many of the problems are correctable and initiatives are already under way to reduce some of them, including registration of protocols and results, and the development of guidelines on good reporting. Study sponsors, public and charitable funding bodies, researchers, and editors still have work to do, however. They could start by listening more carefully to the people they are ultimately working for—patients and the healthcare professionals caring for them.


Cite this as: BMJ 2009;338:b2417