Friday, 19 September 2014

Are antiemetics efficacious in the Emergency Department?

Giving antiemetics in the Emergency Department is a very common ritual that we perform without much thought. But it might be surprising to know that there is very little evidence as to their efficacy. Most studies looking at antiemetics have been performed in patients who are post-operative or in those undergoing chemotherapy.

Ironically, conducting a well-designed trial should be a reasonably simple and feasible undertaking. Nausea is a near ubiquitous complaint in the ED and intuitively it should be easy to create a double blind randomized controlled trial with adequate power to make reasonable conclusions.

These authors sought to compare ondansetron 4mg, metoclopramide 20mg and placebo in adult patients in whom the treating doctor thought they needed an antiemetic. There were lots of sensible exclusion criteria. The primary outcome was a reduction of 30mm on a VAS recorded half an hour after drug administration. There were lots of secondary outcomes investigated.

They ended up with 258 patients for analysis. Mean decrease in VAS score was 27mm for ondansetron, 28mm for metoclopramide, and 23 for placebo. This was not statistically significant. The authors state that this study adds weight to the recommendation that drug use not be routine for nausea and to focus on other specific treatments or supportive care.

However, I am not so sure we can make any conclusions that suggest we throw away our antiemetics. This study was powered to find a 30mm reduction in VAS from baseline. Trends clearly favoured the antiemetics over placebo; therefore a larger study with more statistical power could very likely demonstrate a difference.

The authors seemed to down play the importance of the administration of rescue medications as meaningful outcome measure. In reality, this may actually be the most objective outcome measure as it does not rely on self-reporting of nausea. Many less people in the metoclopramide group required rescue medication.

The limitations section rightfully point out that one could easily argue over the doses of drugs chosen in this study. Why not 8mg of ondansetron and 10mg of metoclopramide?

One important aspect not mentioned in the limitations is the effect of placebo. As I have mentioned before, remember that placebo is not “nothing.” It can have substantial physiologic and psychological effect especially when it comes to pain and nausea. It would have been interesting if they had a “nothing” arm.

Is this study practice changing? Definitely not. But it does raise some interesting questions. A study of many more patients would be necessary to make any meaningful conclusions. I won’t throw away my antiemetics quite yet. But perhaps I am kidding myself.


Egerton-Warburton D, Meek R, Mee MJ, et al. Antiemetic Use for Nausea in Adult Emergency Department Patients: Randomized Controlled Trial Comparing Ondansetron, Metoclopramide, and Placebo. Ann Emerg Med. 2014;?:1-7 [article in press]

Wednesday, 17 September 2014

Oseltamivir (Tamiflu) for Influenza... A Crushing Cochrane Review

Billions of dollars has been spent around the world stockpiling neuraminidase inhibitors to be used in case of pandemic influenza. The presumption is that they are effective, have little side effects, prevent transmission, complications and will save the planet. Was this money well spent?

The 2014 Cochrane review to describe the benefits and harms of oseltamivir was truly a gargantuan and unprecedented undertaking. The authors recognized there was the substantial risk of reporting bias as all of the RCT’s investigating oseltamivir were industry sponsored. Would a drug company that stands to make billions of dollars suppress or hide information that may be damning to their cause? I’ll let you decide.

In order to get around reporting bias, the authors conducted a four year campaign to obtain full clinical study reports of the oseltamivir trial programs. This has never been done before. Full study reports are not simply the papers that get published or rejected from peer review journals. No, they are the raw data and a lot more. In the case of oseltamivir, clinical study reports were of a mean length of approximately 1300 pages each.  In the end, the authors succeeded in accessing the reports and relevant regulatory comments totalling approximately 150,000 pages! How anyone could go through all of this information is beyond me. Heck, I struggled to read their 18 page Cochrane review in the BMJ.

To get right to the point, let’s look at the conclusions.

How well does oseltamivir work? It reduces the time to first alleviation of symptoms by about 16 hours. This represents a reduction in time from having the flu for a full 7 days to about 6.3 days in the treatment group.  This of course likely represents best case scenario as all of the study designs were industry supported.  

Does it reduce admissions to hospital? No.

Could conclusions be drawn on the ability of oseltamivir to interrupt viral transmission? No and nor could the FDA. This leads one to question why this drug has been stockpiled.

What about potential harms of oseltamivir? It increased the risk of nausea (NNTH 28), and vomiting (NNTH 22). In addition it increases the risk of headaches and renal & psychiatric syndromes. One interesting fact to come out of all of this was the “placebo” used in the trials contained dehydrocholic acid and dibasic calcium phosphate dehydrate. Both can cause GI symptoms which may have been a source of bias in favour of minimizing the potential harms of oseltamivir.

The conclusions as stated by the authors strike me as very cautious and measured. It made me wonder if they were concerned about the legal power of a wealthy pharmaceutical company. The last sentence of their abstract simply states, “The trade-off between benefits and harms should be borne in mind when making decisions to use oseltamivir for treatment, prophylaxis, or stockpiling.” It took them 4 years of struggle, expending so much time, effort, and energy to come up with this wimpy conclusion? My guess is the authors would have a different opinion over a glass of beer.

At least the authors did leave a little sting at the end of their manuscript buried on page 9 of 18. They conclude by stating:

Given that oseltamivir is now recommended as an essential medicine for the treatment of seriously ill patients or those in higher risk groups with pandemic influenza, the issues of mode of action, lack of sizeable benefits, and toxicity are of concern. This is made worse by the record and stated intentions of governments to distribute oseltamivir to healthy people to prevent complications and interrupt transmission on the basis of a published evidence base that has been affected by reporting bias, ghost authorship, and poor methods.

We believe these findings provide reason to question the stockpiling of oseltamivir, its inclusion on the WHO list of essential drugs, and its use in clinical practice as an anti-influenza drug.


Jefferson T, Jones M, Doshi P, et al. Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments. BMJ 2014;348:g2545

Wednesday, 10 September 2014

Can We Rule-out MI in Two Hours?

I love any study that provides evidence that allows us to use less time or resources in the Emergency Department. Can I use a high-sensitive troponin (hsTnT) assay to rule-out MI in two hours rather than the traditional six? Yes, but with some cautions and caveats.

The authors of this paper looked to validate an accelerated hsTnT assay measured at 0 and 2 hours for diagnosing acute myocardial infarction in the Emergency Department.

This was a prospective observational cohort study conducted on consecutive eligible patients at the Royal Brisbane and Women’s Hospital in Australia. Inclusion criteria specified adults whereby the treating doctor thought they should perform serial troponin measurements to rule-out MI. In addition to the hsTnT assay they also got all of the usual work-up for potential ACS including ECG’s, “old school” troponin I at 0 & 6 hours and other follow up.

As with all studies investigating a new “better” test, the gold standard is always a little troublesome. To what do you compare your new test against when yours is presumably the best? This study used the usual adjudicated diagnosis by a non-blinded cardiologist based on all of the clinical data available. I do not believe they were blinded to the hsTnT as this was not 100% clear in the text.

What were the results? 764 patients were enrolled. There was a 7.3% (56 patients) rule in MI rate which is rather low compared to usual Australian standards but probably high compared to those in North America. The hsTnT measured at 0 & 2 hours was 96% sensitive and had a negative predictive value of 99%. Only two patients out of 56 were missed with this strategy. Sounds great! But some caution is in order.

The median time from symptom onset to presentation was FIVE hours. Therefore be very careful in patients who present early after chest pain onset. Other studies (Reichlin or Keller NEJM August 27, 2009) have shown that the hsTn assays generally become positive within about three hours. So it is not surprising that the early hsTnT worked well in this study. In fact the sensitivity of the first troponin was 93%. The addition of the two hour troponin only picked up two more patients and increased the sensitivity to 96%.

Don’t get overly excited about the 99% negative predictive value as this of course is a reflection of the low prevalence of disease in this cohort. Remember they started out with a negative predictive value of 93% before any test was performed. This is bound to be worse in a higher risk cohort.

The authors go on to state that time saved by this approach could have a major impact on health service delivery. Not so fast. I agree that two hours is better than six. But the high sensitive troponin assays have already added to our work load with its drop in specificity. More and more patients are “positive” creating a big headache. In the end, I have no doubt they increase further diagnostic testing and admission rates. So I doubt we can consider the high sensitive troponin assays as time or resource saver.

What do I take home from this? If my goal is to rule-out MI, I can probably do this with adequate sensitivity with a 0 and 2 hour  high-sensitive troponin in a patient with reasonably low pre-test clinical suspicion. But not in a patient who is high risk or presents early after chest pain onset. In the end, I am very happy that there is now some more evidence that can support an accelerated rule-out MI strategy in suitable patients.  


Parsonage WA, Greenslade JH, Hammett CJ, et al. Validation of an accelerated high-sensitive troponin T assay protocol in an Australian cohort with chest pain. Med J Aust 2014;200(3) 161-165.

Sunday, 7 September 2014

Paracetamol (Acetaminophen) Does Not Work For Low Back Pain?

This is the message I heard being touted recently by a study author while listening to the Australian ABC news radio. Could this be? Have we been prescribing this medication inappropriately for all these years?

As it turns out, the devil is in the details. It all depends upon which question you are asking.

Enter a large multicentre RCT across 235 primary care centres in Sydney, Australia. 1652 patients with new low back pain were randomized in a 1:1:1 ratio to receive up to 4 weeks of regular paracetamol, as-needed paracetamol or placebo. Patients with suspected serious spinal pathology were excluded from the study. The methods were quite good and the double-dummy blinding was quite clever.

Patients were followed up at 1 week and participants recorded pain scores in to a daily pain and drug diary until recovery or 4 weeks.

The primary outcome was time to recovery in days. Recovery was defined as the first day of 0 or 1 pain intensity, measured on a 0-10 point scale, maintained for 7 consecutive days.

Did I read this correctly? The study authors were trying to determine if paracetamol could cure low back pain? I had no idea that there were those who thought paracetamol could cure low back pain. I was always under the impression that it might take the edge off of the pain a little bit for a brief period of time. But a cure, no.

The authors justified this primary outcome as below.

We chose this outcome on the basis of previous findings of fast recovery in a cohort receiving regular paracetamol. As such, we postulated that regular paracetamol would improve recovery by decreasing pain intensity and allowing people with acute low back pain to remain active and resume normal movement as soon as possible.

The results showed a median time to recovery of 17 days with no statistical difference between the groups. So, paracetamol does not cure low back pain. Big surprise.

To be fair, the study did look at secondary outcomes of pain intensity, disability, function, global rating of symptoms change, sleep quality and quality of life and found no differences. But one could reasonably question the validity of the measuring tools. In addition, one should always be a bit more skeptical when looking at multiple secondary outcomes.

One should also remember that "placebo" is not an inactive treatment arm. To put another way, placebo is not nothing. It can have substantial physiologic and psychological effects in some people especially when it comes to pain. This may have watered down the results of the primary outcome. It would have been interesting if the study had a "nothing" arm. 

In the end, this study truly emphasizes the need for clinicians to be vigilant and emphasizes the value of taking a critical look at the medical literature. If one was not aware of the study’s primary outcome, one might be convinced there is no role for paracetamol for low back pain. This is far from the truth. Nevertheless, I don’t think anyone is deluded that paracetamol does very much. But at least it is cheap, has a low side effect profile and might help for a little bit for a brief period of time.

What’s the take home message? Paracetamol is still fine for low back pain and beware what you hear on the radio.


Williams CM, Maher CG, Latimer J, et al. Efficacy of paracetamol for acute low back pain: a double-blind, randomised controlled trial. Lancet 24, 2014;  [Epub ahead of print]


Williams CM, Latimer J, Maher CG, et al. PACE--the first placebo controlled trial of paracetamol for acute low back pain: design of a randomised controlled trial. BMC Musculoskelet Disord 2010;11:169