Posted by: sayamika, the killer bunny | 2009 July 23

EBM and off-label prescribing

Did you know that taking aspirin immediately after getting home from the beach having forgotten your sunblock can prevent your sunburn from peeling?

Did you know that haloperidol (a powerful antipsychotic and sedative) can really help the nausea associated with various illnesses?

Did you know that several old-fashioned antidepressants are also excellent painkillers?

Traditional medicine, or for the granolas out there, “allopathic” medicine, prides itself on its use of evidence. The evidence in question has several levels, with the highest level of evidence, considered the most reliable, being meta-analysis of multiple randomised controlled trials (RCT). The next level is RCT itself. Then we get down to cohort studies and case-control studies. The final- and lowest- level of evidence is not evidence at all but “expert opinion,” often by consensus at a meeting.

Clinical trials start usually in animals, and once animal trials show some form of desireable effect and good safety, human trials can begin. Human trials have four phases: Phase I is absorption, elimination, tolerability, and safety. Phase II is dosing standards. Phase III are RCTs for efficacy.

Drugs are approved (or not) at Phase III, and can then be put on the market.

Phase IV trials are also known as post-marketing surveillance.

I was recently reading a commentary by Skepdude (of Skepfeeds) regarding the off-label use of antidepressants for behavioural control in autistic youngsters. He (she? it?-ed note- no offense, just don’t know if you are a boy or a girl) was quite upset at the idea that physicians, who talk a good talk about evidence, would merrily go about prescribing drugs for things other than those for which they underwent clinical trials.

Here’s the thing: Phase IV clinical trials often give more information that all of the previous steps combined.

Witness the statins, wherein a lowering of the patient’s cholesterol was used as a surrogate for improving cardiovascular outcomes. If one follows the developments with the statins since the start of the phase IV trials, along with comparisons with fibrates, niacin, and other lipid-lowering drugs, one finds that following the cholesterol as a marker using a statin does correlate with cardiovascular outcomes, but using other drugs, not so much. More recent postmarketing research points to an anti-inflammatory effect from statins, now postulated to be more important than the lipid-lowering effect.

The point here is that all research is based on hypothesis, and hypothesis testing in humans is imperfect at best. Surrogate endpoints are the usual, and are often only associated, not causative, phenomena.

Ranitidine is an antacid, but we use it in treatment of anaphylaxis. Why? Because it is an antihistamine which acts on the digestive tract. It has not been tested for anaphylaxis, nor is it likely to be. It is a reasonable use based on mechanism of action for a known safe drug.

Propranolol is a blood-pressure medication which is also used for migraine prevention, and treatment of benign essential tremor. Again, reasonable uses based on mechanism of action for a known safe drug.

Back to the children with autism. Autism is difficult. It is incredibly stressful to be a parent and caregiver to a child with whom communication is difficult. It can be hard to know the state of mind of children with autism.

The hypothesis: If anxiety causes increased behavioural problems in kids with autism, and antidepressants have been shown to be effective in children with anxiety disorders, antidepressants are likely to help manage behavioural problems in children with autism.

Many behavioural problems of people with autism are attributed to increasing stress levels, as they are seen with changes in the person’s body (ie. adolescence), with illness, or with changes in routine. This may and incorrect assumption.

Certainly, antidepressants can be helpful for children with anxiety disorders who are not autistic. They have risks, but overall, the risk-benefit ratio is quite favourable. The kids are able to express that they feel better, and school grades and participation in fun activities often improve.

Kids with autism are still kids. They may be misinterpreting sensory stimuli. They may have difficulty expressing their needs. They are still people, and one way or another, they do let their caregivers know when they are stressed or anxious. They act out for a variety of reasons.

IMHO, it is reasonable then to consider antidepressant use to reduce anxiety levels in kids with autism, with the aim (for the parents) of making their behaviour more manageable. We don’t know what they are experiencing, but we can reasonably suspect that things that make life easier for other kids may be helpful for them.

Do we know this to be true? Certainly not. But if we do not use drugs off-label, we can never discover if it might be.

It is important that we discuss with patients, caregivers, and parents when a drug is being used off-label. We also need to have the usual “these are the common side effects and these are the dangerous ones” discussion. Informed consent demands no less.

Sometimes a new population can have unforseen risks. This can be the case with any drug: trials are done on very defined subpopulations which may or may not represent the actual population who will eventually get the drug in question. Jupiter, for example, measures statin effect on CRP in patients with no inflammatory comorbidities. Not exactly target market, really.

Phase IV clinical trials, or post-marketing research, gather information on off-label use for potential further RCTs (can you say possible patent extension and an excuse for rebranding?), as well as information regarding heretofore unknown side effects (I won’t say Vioxx, but the phase IV trials made it impossible to keep a lid on that). You can call it “experimenting on kids,” but that’s an oversimplification.

For some populations, all medication use is off-label. Morphine is a great painkiller, but small doses also relieve the air-hungry feeling of people dying of lung disease. Remeron increases the appetite in patients with AIDS cachexia, and gives them a chance to put a little weight back on. There is no impetus to do research on such effects, so these uses, though well-known, remain off-label.

Off-label use is a part of the art of medicine. Evidence-based medicine is powerful, but should not turn us into protocol-driven automatons. Clinical judgement and an understanding of physiology, pharmacology, and common sense are part of who we are as physicians.



  1. It’s a he.

    I just think that the standards of evidence should be the same. Just a little while ago all cold medication for kids under 2 was pulled, because it was being prescribed without being tested on kids. Now the same happened with this drug that was being prescribed for autism.

    It is true that I am an outsider to the medical world and do not understand how all the wheels turn, but it seems to me that the burden of proof is the same if we’re dealing with woo meisters or doctors.

  2. You’re not wrong. In an ideal world, that’s how it would be.

    It’s just not how things are. Think about pregnancy: once upon a time, it was considered OK to test drugs on pregnant women, now it’s not.

    There is a total of one drug tested on pregnant women for pregnancy induced nausea. Excellent. It’s extortionately priced, which means if you’re broke and pregnant, have fun with Ralph.

    Antidepressants cannot be tested on pregnant women. I am suicidal without my antidepressant. WTF do I do when I get pregnant?

    Answer: weigh the risks and benefits as they stand today, and report to phase IV marketing so that my experience can make the decision easier for someone in the future.

    The reason I posted on my own blog was my thought ran on longer than was seemly for a reply. I linked in the hopes that wordpress would tell you about the link, but I didn’t feel like I should go, hey, look over here! Anyhow, thanks for the reply.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: