Research News

All you need to read in the other general journals

BMJ 2012; 344 doi: (Published 23 May 2012) Cite this as: BMJ 2012;344:e3578

FDA approves new drugs faster than European or Canadian regulators

Regulatory agencies are often accused of being slow to approve new drugs, and the US Food and Drug Administration (FDA) is under pressure to speed up ahead of a five yearly review of its charges. The pressure may be misplaced, say researchers, who report that the FDA approves new drugs significantly faster than both Canadian and European regulators.

In a head to head comparison of review times for new drugs approved between 2000 and 2010, the FDA took a median of 322 days to complete review, compared with 366 days for the European Medicines Agency and 393 days for Health Canada. The researchers analysed 510 approvals of 289 individual drugs, using publicly available data from all three regulators. Nearly two thirds of the drugs approved in both the US and Europe were approved by the FDA first (121/190). More than four fifths of the drugs approved in both the US and Canada were approved by the FDA first (132/154).

The researchers weren’t able to look at review times for unapproved drugs because these records are not accessible to the public. Neither did they examine the quality of regulators’ decisions. Rapid approval must be balanced against public safety, they write. Future studies should compare the regulators for black box warnings, label changes, drug withdrawals, and other possible indicators of hasty and ineffective decision making.

Genetic study challenges causal link between HDL-C and risk of heart attack

The association between higher serum concentrations of high density lipoprotein cholesterol (HDL-C) and a lower risk of heart attack is well known. It now seems unlikely that the link is causal, so developing drugs to raise HDL-C concentrations could be fruitless. An international team of researchers has found that adults who are genetically programmed to have high HDL-C have the same risk of heart attack as everyone else (odds ratio 0.99, 95% CI 0.88 to 1.11). They would have had a 13% lower risk if HDL-C cholesterol had been directly protective and not just a biomarker, say the researchers.

The new study exploited nature’s random allocation of genes to investigate the association between HDL-C and heart attack in more than 100 000 adults. Genetically high HDL-C cholesterol did not seem to protect against heart attack in either cohort or case-control analyses. The researchers studied one rare but powerful allele that increases HDL-C (LIPG 396Ser allele) and a panel of 14 less powerful alleles. The results were the same, and they give the strongest indication yet that interventions to raise HDL-C cholesterol are unlikely to prevent heart attacks directly.

These findings are consistent with previous smaller Mendelian randomisation studies and also with clinical experience of rare inherited disorders of HDL-C metabolism, says a linked comment (doi:10.1016/S0140-6736(12)60481-4). People with Tangier disease, for example, don’t develop early coronary artery disease despite very low concentrations of HDL-C.

Low risk adults benefit from statins too

Statins help prevent major vascular events in people with and without cardiovascular disease, although international guidelines generally recommend treatment only for people whose predicted risk of a vascular event exceeds at least 10%, or even 20%. The latest meta-analysis of statin trials suggests that adults below typical thresholds have as much to gain and as little to lose from statin treatment as those above these thresholds.

Among adults with a five year predicted risk of just 5-10%, statins reduced major vascular events by 30% for every 1 mm/L drop in low density lipoprotein (LDL) cholesterol concentration (relative risk 0.69, 99% CI 0.60 to 0.79). This reduction was driven mainly by fewer non-fatal heart attacks and coronary revascularisations. Guidelines should be revisited, say the authors. They estimate that treating 1000 adults with a 5-10% baseline risk would prevent an extra 11 major vascular events over five years, including four deaths from vascular events.

Treating low risk adults with statins looks effective, says a linked comment (doi:10.1016/S0140-6736(12)60694-1), but is it feasible, or even desirable? In the UK, half of all men over 50 years are already above the current risk threshold for treatment with statins. Lowering the threshold, even to 10%, would mean treating 83% of men over 50 and more than half of all women over 60. On current performance, primary care would simply fail to cope, even if such whole sale medication was affordable and healthy symptomless adults could be persuaded to take their tablets for half a lifetime (both big “ifs”).

Alternative approaches should not be overlooked, says the comment. When the government of Mauritius switched their oil import policy from palm oil to soya oil, mean cholesterol concentrations fell by 0.8 mmol/L, for free.

Flexible sigmoidoscopy screening prevents cancers and saves lives

A large randomised trial from the US has confirmed that screening with flexible sigmoidoscopy reduces mortality from colorectal cancer (2.9 deaths per 10 000 person years in the screened group v 3.9 in the control group; relative risk 0.74, 95% CI 0.63 to 0.87). The trial compared two rounds of screening three or five years apart with usual care in almost 155 000 adults over 55 years. The screened group had a lower incidence of both distal and proximal cancers than controls during 12 years of follow-up. The mortality benefit was confined to deaths from cancers distal to the splenic flexure.

Most adults with positive findings at sigmoidoscopy went on to have a diagnostic or therapeutic colonoscopy. A fifth of men and 13% of women in the trial had a false positive sigmoidoscopy. Their risk of perforation during subsequent colonoscopy was 107.5 per 100 000 procedures.

Flexible sigmoidoscopy clearly works as a screening test for colorectal cancer, says a linked editorial (doi:10.1056/nejme1204099). We now have good evidence from the US and Europe that this type of screening can prevent deaths, at least from cancers in the distal colon. Whether colonoscopy can do any better and prevent deaths from proximal cancers too remains unknown. Proximal lesions are flatter, harder to see, trickier to remove, and biologically distinct from more distal lesions. There are still no randomised trials evaluating colonoscopy.

Flexible sigmoidoscopy is falling out of favour in the US, says the editorial. The decline can’t be justified on current evidence, and there are good reasons to reverse it. Patient preference for less invasive screening options is one of them.

End of the line for drotrecogin alfa in septic shock

There have been many false dawns in the search for lifesaving treatments for severe sepsis. Such treatments include drotrecogin alfa, a controversial agent that seemed in an early trial to cut deaths by about a fifth. Regulators in the US and Europe granted limited licences a decade ago and demanded further trials. All failed to confirm the drug’s initial promise. The final trial reported this week that drotrecogin alfa works no better than placebo in adults who are critically ill with sepsis, shock, and hypoperfusion. A quarter of both groups died within 28 days of randomisation (relative risk in the treated group 1.09, 95% CI 0.92 to 1.28). Drotrecogin alfa did not prevent deaths at any time point or in any subgroup. The authors can’t explain the discrepant results between their last trial and the first.

A linked editorial (doi:10.1056/nejme1203412) believes that this marks the end of the line for drotrecogin alfa and the closure of another chapter in a history littered with failed attempts to modify the complex inflammatory response to infection by blocking individual triggers, such as tumour necrosis factors and interleukins. Drotrecogin alfa is a recombinant human activated protein C, which helps reverse the microvascular coagulation associated with sepsis, so an effect on mortality was always biologically plausible. The theory has not translated into fewer deaths in practice, says the editorial, and researchers will once again have to go back to the biology of sepsis for further inspiration.

Multicentre ethics review causes delays and spiralling costs, even for low risk trials

OpenUrlAbstract/FREE Full Text

Ethical review of multicentre trials can be slow and difficult. In one case study from the US, a multicentre protocol to test the effect of financial incentives on evidence based prescribing took 827 days to work through the requirements of dozens of different institutional review boards. The whole exercise cost $168 229 (£106<thin>354; €132<thin>228) in staff salaries alone. Twenty five sites were eligible for the trial, but only 17 had the staff and resources to embark on ethical review. Fourteen sites were eventually approved (after a median of six submissions each) and just 12 recruited any patients. They were not typical of eligible sites, and external validity was compromised, say the authors. Review took 21 months longer than researchers had planned or budgeted for.

Their trial was judged to be “minimal risk” health services research, and these experiences confirm that idiosyncratic and inefficient review is a problem across the board, not just for clinical trials. The system of piecemeal review operating in the US can work against the interest of patients and of society, says a linked editorial (p 746). It must be reformed. Central review of multicentre trials is one option. Another is to make better use of data monitoring committees, which already make difficult ethical judgments when suggestions of harm emerge from interim analyses.

The pros and cons of lactobacilli for recurrent UTI


Capsules of Lactobacillus spp, taken orally, were not as good as antibiotics at preventing recurrent urinary tract infections in a recent trial. But they had one important advantage, say researchers. Lactobacilli did not induce antibiotic resistance. Most isolates of Escherichia coli from women taking trimethoprim sulfamethoxazole were resistant to the antibiotic within a month. All isolates from urine were resistant within 12 months.

Women in the trial were postmenopausal and reported a mean of seven urinary tract infections in the year before recruitment. They took capsules of trimethoprim sulfamethoxazole (480 mg) or lactobacilli (L rhamnosus GR-1 and L reuteri RC-14) for 12 months. Those taking antibiotics had fewer clinical recurrences during treatment (2.9 v 3.3; mean difference 0.4, 95% CI −0.4 to 1.5). The difference wasn’t significant but it was big enough to rule out a “non-inferior” result. Women taking antibiotics had significantly fewer microbiological recurrences (1.2 v 1.8; P=0.02).

Lactobacilli may be a second best treatment in this setting, but they cause substantially less collateral damage than antibiotics, says a linked comment (p 712). Avoiding collateral damage to normal human flora is fast becoming the more important outcome.


Cite this as: BMJ 2012;344:e3578