The month in research
Early stomach cancer
Eradicate H pylori after removal
Lancet 2008;372:392-7
Prophylactic eradication of Helicobacter pylori seems to be effective in the prevention of monochronous gastric cancer after endoscopic removal of early stomach cancer.
A Japanese trial looked at 544 people who had either undergone or were scheduled to undergo endoscopic removal of early gastric
cancer and were thus at high risk of developing monochronous cancer. These patients were randomised to receive H pylori eradication treatment (lansoprazole 30 mg twice daily, amoxicillin 750 mg twice daily, and clarithromycin 200 mg twice daily
for a week) or usual care that did not include antibiotics.1
After three years, metachronous gastric cancer developed in nine people randomised to the eradication group and in 24 people
who received usual care (odds ratio 0.35, 95% CI 0.16 to 0.78). The only notable adverse events seen in the eradication group
were soft stools (12% (32/255)) and diarrhoea (7% (19/255)). A commentator (p 350) argues that, in the absence of large trials
of primary prevention, the implications of these findings should extend beyond high risk patients, and that eradication of
H pylori to prevent stomach cancer should become a priority in regions with high incidence. Colorectal cancer kills fewer people,
says the comment, yet screening by colonoscopy is widely accepted, despite the lack of direct evidence of benefits and the
risks of such a strategy. (This item was first published in the BMJ, 2008;337:a1129.)
Glucose control
May harm critically ill patients
JAMA 2008;300:933-44
In 2001, a trial found that keeping blood glucose below 8.3 mmol/l or 6.1 mmol/l (moderately tight and very tight glucose
control, respectively) by infusing insulin during part or all of the stay in intensive care reduced in-hospital mortality
by one third in critically ill surgical patients. Since then, tight glucose control in all critically ill patients has been
recommended and endorsed by a growing number of guidelines and professional societies. However, later studies have largely
failed to replicate the original finding.
A systematic review and meta-analysis based on 29 randomised trials and nearly 8500 critically ill patients now shows that,
compared with usual care, tight glucose control does not improve survival during hospital stay and does not reduce the need
for new onset dialysis, although it reduces the risk of septicaemia (10.9% v 13.4%; relative risk 0.76, 95% confidence interval 0.59 to 0.97), but only for patients in surgical intensive care units.
Most notably, hypoglycaemia—which is associated with serious neurological events in critically ill patients, ranging from
seizures to coma—is five times more common with tight glucose control than with usual care (13.7% v 2.5%; 5.13, 4.09 to 6.43).
The editorialists (p 963) say that these findings may surprise many clinicians, but it is still unclear whether tight glucose
control in critically ill patients is right or wrong. Standards and definitions need to be universally agreed upon, and the
hunt is still on for the optimal conditions in which tight glucose control might confer benefits to patients. (This item was
first published in the BMJ, 2008;337:a1510.)
Type 2 diabetes
Even small doses of arsenic raise risk
JAMA 2008;300:814-22
Exposure to high doses of inorganic arsenic, which is mostly ingested through drinking water, is known to increase the risk
of type 2 diabetes mellitus, but few studies have looked at exposure to smaller doses. Arsenobetaine—an organic arsenic compound
mainly ingested through seafood—is excreted unchanged in urine and considered non-toxic, but this has not been tested in human
studies.
A population based health and nutrition survey (NHANES 2003-4) measured urine concentrations of arsenic compounds in 788 US
adults. After adjustment for diabetes risk factors and seafood intake, people with type 2 diabetes had 26% (95% confidence
interval 2.0% to 56%) higher concentrations of total urine arsenic than people without diabetes, but concentrations of organic
arsenic did not differ in the two groups, which confirms the notion that this type of arsenic is non-toxic in humans. When
compared with people with the lowest total urinary arsenic concentrations, people with the highest concentrations had a 3.6-fold
higher risk of type 2 diabetes.
The exposure to arsenic seen in the study may have been as little as three times less than the current US Environmental Protection
Agency reference dose of 0.3 μg/kg body weight each day, say the editorialists (p 845). The association needs to be explored
further in prospective studies, but in the meantime it seems prudent to minimise exposure to arsenic. (This item was first
published in the BMJ, 2008;337:a1423.)
Childhood pneumonia
Time to reconsider guidelines?
Lancet 2008, doi: 10.1016/S0140-6736(08)61166-6
Guidelines recommend that children with severe pneumonia are referred to hospital. However, in some settings, geographical,
financial, or cultural barriers may result in children not reaching hospital. In Matlab, rural Bangladesh, a before and after
study explored changes in care when guidelines were modified so that most of these children (aged 2 months to 5 years) were
treated with oral amoxicillin at home, prescribed by local first level facilities, with referral to hospital reserved for
those with danger signs or severe symptoms. Before the modified guidelines were implemented, most children with severe pneumonia
were referred to a hospital (245/261 (94%)). However, only a third (94/261) of children received appropriate treatment, and
1.1% of children died. After the guidelines were modified, only 8% (107/1271) of children were referred, but 90% (1145/1271)
received correct treatment and the case fatality rate dropped to 0.6%.
The commentators (doi:10.1016/S0140-6736(08)61167-8) think that the existing guidelines should be simplified by combining current classifications of pneumonia and severe pneumonia
into one episode of pneumonia, and by treating all children with pneumonia at home with oral amoxicillin. Such a strategy
needs to be tested further, but it is likely to be widely successful, except in settings with a high burden of HIV. (This
item was first published in the BMJ, 2008;337:a1423.)
Cerebral palsy
Magnesium sulphate in premature infants
N Engl J Med 2008;359:895-905
Magnesium sulphate has been widely used to prevent uterine contractions, even though it has been shown to be ineffective,
but it is indicated for the prevention and treatment of pre-eclampsia. Several observational studies have indicated that it
might prevent neurological birth disorders.
A cluster randomised double blind placebo controlled trial recruited 2241 pregnant women who were at high risk of premature
labour. Magnesium sulphate, given intravenously as a 6 g bolus followed by a constant infusion of 2 g per hour, was no different
to placebo in its effect on the primary outcome—the composite of stillbirth or infant death by 1 year of age or moderate or
severe cerebral palsy at or beyond 2 years of age. However, prespecified secondary analyses showed that magnesium sulfate
reduced the occurrence of moderate to severe cerebral palsy (20/1041 (1.9%) v 38/1095 (3.5%), relative risk 0.55, 95% confidence interval 0.32 to 0.95) and also reduced overall occurrence of cerebral
palsy compared with placebo (40/942 (4.2%) v 73/1002 (7.3%)).
The editorialists (p 962) say that these results look promising, but before this preventive treatment is widely recommended
we need a better understanding of the factors that might influence its success, such as the reason for premature birth, the
timing of treatment, and the amount of drug given. (This item was first published in the BMJ, 2008;337:a1510.)
Student BMJ 2008;16:377 | 10
