more_vert
NEW DRUGS have to jump through all sorts of hoops to get approval for marketing in Australia, but once they’re out there, it’s a different story.
Experts have been calling for a more coordinated system for monitoring drug safety in Australia for years – and particularly since the September 2004 withdrawal of the selective COX-2 inhibitor Vioxx over concerns about cardiovascular risk – but we’re still waiting.
The cardiovascular safety of NSAIDs was back in the news last week after a large Danish observational study conducted in healthy individuals found that various NSAIDs were associated with an increased stroke risk – from 28% for ibuprofen to 86% for diclofenac.
The study correlated NSAID use and stroke data in half a million people aged over 10 years, controlled for age, sex and socioeconomic status. Those with chronic diseases or a hospital admission within the previous 5 years were excluded.
Cardiologist Gunnar Gislason, who presented the data at the recent European Society of Cardiology Congress, called for NSAIDs to be available only on prescription in light of the findings.
It’s no accident this new data came from Scandinavia, where postmarketing surveillance systems allow linkage of prescribing data to other information, including hospital admissions and mortality figures.
If similar effects are associated with the drugs here, we quite simply have no way of knowing about them.
Observational studies like this latest Danish one obviously don’t prove cause and effect, but they can provide an effective warning system for real-world adverse effects that would be unlikely to show up in premarketing clinical trials.
You don’t see many drug trials conducted in elderly people with multiple comorbidities, for example, yet those are precisely the people who are likely to be the biggest users of most pharmaceuticals – and the ones at most risk of adverse effects.
Safety data for these groups generally only become available after the drugs hit the market, and we, in Australia, basically don’t bother collecting it.
Yes, there are potential privacy concerns when you start talking about linking data from various parts of the health system, but they’re not insurmountable.
Surely it’s time we paid proper attention to the safety of medications after they’re released into the general population as well as before?
Jane McCredie is a Sydney-based science and medicine writer. She has worked for Melbourne’s The Age and contributed to publications including the BMJ, The Australian and the Sydney Morning Herald. She is also a former news and features editor with Australian Doctor magazine. Her book, The sex factory, on the science of sex and gender will be published by UNSW Press later this year.
Posted 20 September 2010
By using publicly available clinical trial data and linked de-identified hospital morbidity and PBS medications dispensing data it is possible to develop an age- and gender-sorted risk profile for current Australian patient groups who have the morbidity targeted by a new drug, but who are likely to be at risk of an adverse drug event if they are prescribed that new drug in the context of their age, gender, comorbidities and co-medications.
This risk profile can be developed without waiting for the accumulation of data in post-marketing surveillance to identify areas of exposure to ADE risk.
A risk profile developed for arthritis patients who were aged over 65 before rofecoxib was available via the PBS showed that 68% of each of two annual cohorts carried a cardiovascular risk based on a hospital morbidity diagnosis, or prescriptions for CVD medications or both.
Given that the rofecoxib clinical trials had exclusion protocols in place which strictly limited the participation of patients with arthritis who also were receiving CVD treatment, availability of a risk profile and associated warnings could have reduced the adverse CVD outcomes for many Australians who were prescribed rofecoxib.
To say that post-marketing surveillance is merely important is I think an understatement. Crucial would be more appropriate. Achieving a ‘balance’ in the risk-benefit equation is only possible if accurate information is available on both. Currently, as Jane McCredie points out, such information is not available.
Evidence from real-life experience in the use of medicines – both positive and negative – is only accumulated over time and by observing large numbers of patients. This level of evidence is rarely achieved prior to marketing, but can be by utilising population-level observational data.
Observational data describes real patients; some of whom have multiple co-morbidities and take multiple medicines. Such information is included in the data sets already collected for the national death index, PBS, MBS and hospital administration.
Careful case definition and analysis are of course essential. Furthermore, these data sets could be improved and fine-tuned to specifically support outcomes monitoring. However, Australia already has exceptional electronic resources that when confidentially linked, can provide valuable insights into health care outcomes – especially the benefits and risks of community medicine use.
It is important that such routine surveillance be conducted by an independent authority. With rofecoxib we have seen only recently how adverse events data can be withheld or misconstrued by those with vested interests. This fiasco was a disaster for thousands of patients and also for manufacturer Merck. It would seem that early detection of adverse events by a formal medicines monitoring system could be a win-win situation for both industry and the public.
I suspect that in future times, we will look back in amazement that for so long we considered it acceptable to ignore the obvious need for ongoing routine assessment of treatment outcomes. Such need can readily be justified using arguments based on economics, efficiency or ethics – or all of the above.
Yes, post-marketing surveillance is important, but it’s also important to understand the nature and magnitude of risk, and not to throw the baby out with the bathwater.
Firstly, in considering the study, we need to know what the authors mean by “stroke”. Does it include both haemorrhagic and thromboembolic stroke? If anyone who had been in hospital in the previous five years was excluded, what was the rate of stroke in those not taking NSAIDs? If it is vanishingly small, does increasing it by 28% produce a significant number?
Secondly, if we want to be able to apply study data to real life patient populations, we should be looking at NSAID effects in patients who generally take them – older people who DO have comorbidities. We already know that NSAIDs can cause sodium and fluid retention, gastric ulceration and renal complications, but these risks are balanced against the symptomatic benefits.
The real paradox is that, the better we get at testing drugs and looking for complications, the more worried we get. There are few aspects of life without risk – it’s all about balance.
I am a strong supporter of post-marketing surveillance of all medicines. Australia has exceptional data resources that could support a comprehensive drug-monitoring system. I have summarised my views in a co-authored 2007 MJA paper:
“Evaluating medicines: let’s use all the evidence “
http://www.mja.com.au/public/issues/186_05_050307/kel10659_fm.html
Also, see the linked editorial “Australia needs a better system for health care evaluation”:
http://www.mja.com.au/public/issues/186_05_050307/sta11308_fm.html