Black Box Lowered Antipsychotic Use In The Elderly
14 CommentsBy Ed Silverman // January 12th, 2010 // 1:28 pm
The use of atypical antipsychotics to treat elderly patients with dementia appears to have decreased following an advisory issued in 2005 by the FDA concerning the increased risk of death, according to a report in the Archives of Internal Medicine, one of the JAMA/Archives journals. This was an off-label use, by the way.
Researchers analyzed nationally representative data to assess rates of atypical antipsychotic drug use between January 2003 and December 2008. From January 2003 to March 2005, the rate of atypical drug mentions increased 34 percent per year, including a 16 percent increase among patients with dementia. In the year before the FDA advisory, approximately 13.6 million atypical antipsychotic mentions occurred, 0.8 million of which involved patients with dementia (here is the abstract).
An overall decline in the use of atypicals began within one month of the FDA advisory. Mentions of atypical antipsychotics decreased 2 percent overall and 19 percent among those with dementia in the year following the warning. By 2008, monthly drug usage among elderly patients with dementia decreased more than 50 percent, according to the study. Just the same, atypical antipsychotics still comprised 9 percent of prescription drug uses for dementia among older adults by the end of 2008.
“The residual use in the population at risk and the decrease in the use of atypical antipsychotics in the general population, who were not targeted by the warning, raise the question as to whether the effect and specificity of FDA regulatory actions could be enhanced,” the authors wrote. “Targeting specific segments of patients and physicians and further customizing and evaluating the impact of regulatory actions may improve their impact at minimizing the risks associated with select prescription medications.”
Author Affiliations: Department of Neurology, University of Rochester Medical Center, Rochester, New York (Dr Dorsey); Section of General Internal Medicine, Department of Medicine (Drs Rabbani and Alexander and Ms Gallagher), Center for Health and the Social Sciences (Drs Rabbani, Conti, and Alexander), Section of Hematology/Oncology, Department of Pediatrics (Dr Conti), and MacLean Center for Clinical Medical Ethics (Dr Alexander), University of Chicago, Chicago, Illinois; and Department of Pharmacy Practice, University of Illinois at Chicago School of Pharmacy, Chicago (Dr Alexander).
Lisa Van Syckel
Unfortunately we havent seen a dip in the prescribing of antipsychotics in children for ADD/ADHD, or any other condition for that matter.
FDAer
It’s common knowledge and frequently discussed between FDA staff and management that we know that FDA warnings and risk communications do little to decrease inappropriate prescribing. We suspect that drops in usage may be due to decreased usage among those patients who are more risk adverse.
I and other reviewers know that many times management ignore early warning signals, then continue to dismiss and downplay additonal signals. We suspect it’s in order to allow early uptake and adoption of usage which then results in inadequate drops when warnings do come out later. In fact we see the same so called internal experts repeated brought in and they always find ways to dismiss or a left introduce doubt into if AEs are drug related. Yet in my experience in every case these ‘experts’ have always been eventually proven to be wrong. In some cases the signals are the same as for other drugs they were involved in previously which had to eventually be taken of the market and the new jr. reviewer who identified the signal with the new drug was unaware of this.
pharmavet
Maybe we’ll see less of the standing nursing home order for “five at five”. This refers to the standing order on the charts of agitated elderly dementia patients to receive 5 mg of Zyprexa every evening at 5 PM so that they will remain calm (i.e. zonked out) throughout the evening, so that the staff can avoid doing their jobs.
M Helm, MD
Lisa Van S.
You are correct, and in fact, the view from where I sit is that the newer antipsychotics have continued their dramatic increase in use among children (even preschoolers). There appear to be many reasons for this, but the most important seems to be a lack of creativity on the provider’s part to come up with another plan for identifying the real source of disruptive behaviors and developing an effective plan whicn involves changing family dynamics or diagnosing and treating actual underlying problems. A second important cause is that parents seek these medications to give to their children.
A challenge for providers (and possible explanation of why more Medicaid kids are treated with the medications) is that reimbursement for time spent and services delivered to these children and their families is typically dismal. It is hard to cover the costs of the business if you are spending an hour with a child and family with a signficant problem if this will generate less than the costs of keeping the doors open. This produces a strong incentive to treat children with behavioral problems with the same high volume, minimal therapy (maximum reliance on medication) approaches which are now the norm in adult psychiatry. To my understanding, these approaches haven’t significantly improved care for the seriously mentally ill - whereas community support programs seem to make a difference. The adult model of medication management visits seems particularly inappropriate for children - especially children who are often growing up in very high-risk environments.
In my state the last data from a few years ago put the rate of use of newer antipsychotics at 4% of the school age population. ADHD medications were used at between 8% to 12%. In the past couple of years, I’ve increasingly seen both medicines used together.
I believe this reflects a real ignorance of the pharmacology of these medications. Stimulants are known to increase dopamine and norepinephrine neurotransmitter release. The newer antipsychotics block (or modulate) the recptors (or actions) of dopamine and other neurotransmiters. What ends up happening is pharmacologic tail-chasing with the child being subjected to ever-increasing (and increasingly toxic) doses of both medications. The answer to the medication dilema presented is often to (carefully) withdraw both medicines and determine the precise cause of the behavior.
Sometimes the real answer is that the parent has unrealistic expectations of child behavior. True story as a case in point: I know of a provider requesting authorization for coverage of a newer antipsychotic for a child between the ages of 2 and 3. The reason for the request was that the parents reported the child required constant adult supervision or would wander of and/or engage in dangerous activities. Until only a few months ago in our state, this provider would not have been required to request authorization for this child at all.
I’m going to acknowledge that there are some children who need and benefit from some of the psychiatric medications, including newer antipsychotics. However, I don’t think it is 1 in 25 children. I fear the only thing that will change the pattern of use of these potentially dangerous and harmful medicine in our children is the threat of a tidal wave of malpractice claims against providers who prescribe them for experimental/off-label uses.
M Helm, MD
Ed,
Just an editorial suggestion: “Atypical Antipsychotic” as a label is actually a marketing term. With the exception of clozapine, the newer antipsychotic medications are very similiar to the older medications. The notable exception is an increased risk of metabolic side effects in many of the newer agents. I hope you agree that the labeling of something is important. I worry that to call these medicines “atypical” rather than “newer” is to perpetuate a suggestion or perception that these agents as a “class” have special advantage beyond simply being newer and more expensive (and more profitable for their manufacturers).
I’m aware that journal publications often use the term “atypical,” however, this too may be a result PhRMA marketing efforts/pressures.
Humbly Submitted.
riv
I have learned anti-psychotics are referred to in care centres here as “sleeping tablets”. That’s the lingo the elderly are given, and it’s accepted, because sleep is a good thing isn’t it? It eventually gets shortened to just “tablets”. Tablets they take at meal time (the rest) and tablets they take at bedtime. I recognized zyprexa, xanax, seroquel, ativan in the junk three over-80 women whom I try to help. Family is far away. I cannot change anything. Family says “the experts know best”. The person in care pleads that I do something.
So there Dr. Helm. I’ve one-upped you in euphemistic under-the-rug vernacular. Tablets.
patrons99
Even “sleeping pills” may not always be a good thing. Just ask Tiger Woods. I’d strongly argue that sleep aids, in general, do not really serve a vital public health need. Most are just “me-too” drugs, that only serve to deplete healthcare budgets worldwide. Many of them are actually dangerous. As for Dr Helm’s editorial suggestion, I’m inclined to agree with him, although I must say I’m not a specialist in this area. Use of the phrase “Newer antipsychotics” instead of “atypical antipsychotics” seems logical, if they belong to the same drug class.
pharmavet
Dr Helm, the term “atypical” is not a marketing term. The atypicals work for both the negative symptoms of schizophrenia (e.g. social withdrawal, poverty of speech) as well as the positive symptoms (i.e. delusions, hallucinations). They also have a lower predilection for extrapyramidal symptoms (tardive dyskinesia) compared to older neuroleptics. Metabolic side effects are similar between older and newer agents.
Salmon
Actually there is no real difference in the effects of ‘atypicals’ vs. ‘classical’ antipsychotics in effects on either positive or negative symptoms. The main driver in the effect on the PANSS (Positive and Negative Symptom Score) is the change in the positive score scale with both classes of drugs. In some studies there appears to be a slight 1 - 1.5 point greater improvement in negative scores with ‘atypical’s but the improvement in negative scores between the 2 types are not statistically significant. Even more importantly the difference in the negative symptom subscale is not statistically different from the effect of placebo.
The differences in EPS are also likely overblown. There are clinical differences and the ‘atypicals’ are better (10% vs. 15%-20%) but the it’s not as large as generally claimed as rates are often based on excessive doses of classical antipsychotics and relatively low doses of atypicals. When the higher doses used in bipolar are considered the difference in rates largely disappears. Plus the good head to head rates are based on short term acute studies and so they don’t include tardive dyskinesias. Rather you have to infer a difference based on any differences in rates with other types of EPS (i.e. pseudoparkinsonism, akathesia) and of course we know that certain of the ‘atypicals’ such as aripiprazole (Abilify - BMS) have incredible amounts of akathesia.
As for clozapine vs. other azapine and even bicyclic heterocycle type ‘atypicals’. It is an incredibly nonspecific and dirty drug necessitating extremely high dosages which likely contributes to its high rate of agranulocytosis. (The only reason it’s lower with Zyprexa, Seroquel and now Saphris is they’re more potent and thus there isn’t the same molar dose exposure to the toxic N-oxide metabolites.) However they all hit various serotonin receptors and the only reason clozapine has probably shown efficacy in subjects who didn’t respond to the classical antipsychotics is we were selecting out a small subpopulation who didn’t respond to dopamine blockade but did respond to something different and thus nothing had worked previously. However we would likely have seen the same thing if one of the other atypicals had come to market first. So there is likely small subpops who respond to the atypicals who don’t respond to the classical and visa versa and as always there will patients who respond to one atypical and not another. But there is no data that clozapine is anything special (with regards to efficatiousness) other than the fact of when it was developed and that it was the first in a new class of molecules with slightly different pharmacology.
If you look at the PET data for clozapine and the other atypicals clinical dosing actually matches up with the degree of dopamine blockade and is similar to what is seen with the classical antipsychotics. As I recall the mistake that was made was that the PET studies with clozapine purporting to showing greater efficacy with lower dopamine binding was that it was conducted in those subjects who had been refractory to dopamine blockers and yet responded to this new class. Thus it was a biased study that selected out a specific subpop.
Salmon
patrons99
Here’s a naive question: has anyone studied the incidence of various neuropsychiatric (e.g. ADHD, schizophrenia, bipolar disorder) and metabolic conditions (e.g. diabetes) over time for various age groups? If disease prevalence is increasing over time, shouldn’t we ask why? Are there environmental factors at work? At the risk of offending the pro-vaccinators in our group, should we also reasonably ask whether some of these conditions might be “vaccinoses”?
M Helm, MD
Thanks for the back-up Salmon. I would suggest that Pharmavet may wish to read the TEOSS and various CATIE study results.
As I understand it, negative symptoms are a significant unsolved problem in treatment of schizophrenia. I understand that various augmentation approaches have been tried including pharmacologic (selegiline, mirtazapine and minoncycline) and interpersonal/cognitive (more trials from Europe than the US - not surprisingly).
Even if Pharmavet were correct, treatment of schizophrenia still accounts for only a portion of the revenue generated by these medicattions. A very large portion of revenue for these medications is from the experimental treatment of conditions and patient populations in which there is inadequate clinical data. These uses are justified by “clinical judgement.” As stated before, I acknowledge that these medications may offer some benefits to some patients. However, In my reading of many scores of studies, I can not conclude that the newer agents offer substantial advantages. When compared to older agents, and even behavioral interventions, the older agents (or a behavioral intervention) tend to look like better options - particularly when evaluated against the marketing claims such as those echoed by Pharmavet.
Nevertheless, if Pharmavet can cite references proving the assertions the forum would be enlightened, and I’d be delighted to read (or more likely revisit) them. Problem with most of the trials of all of these agents is that a change in mean PANSS ratings (especially on the negative symptom score subsets) may be statistically significant, but not clinically meaningful or important. It is little improvement to go from horrible to miserable, or from miserable to dismal.
Patrons99 poses an interesting question. The databases needed to answer such a question (at least for the medical/pharmacologic interventions) could only exist in an environment with comprehensive data collectin from birth. That can be true in some countries with national health plans. However, there are other environmental factors which could not be controlled for. It is fair to wonder about vaccines, but I wonder more about the effect of the American diet (food that is an abstraction of what is found in nature delivered). Just considering the metabolic and biochemical effects of high fructose corn syrup consumption is frightening. But again, I’m off-topic…
The question that Riv brings us back to is: If we know that these medications increase the risk of death, why do “care-providers” still prescribe and adminster ANY “tablets” to the vulnerable population for whom they are responsible?
Salmon
Dr. Helm,
I agree that the change in Negative scores on the PANSS is likely not clinically significant.
The following is PANSS data from the FDA reviews of the latest antipsychotic approved
Baseline Mean (SE)
PBO 89.0 (0.9)
Asenapine 5 mg BID 88.9 (1.0)
Asenapine 10 mg BID 89.4 (1.0)
Haloperidol 4 mg BID 88.5 (1.0)
Endpoint
PBO Asenapine 5 mg BID 88.9 (1.0)
Asenapine 10 mg BID 89.4 (1.0)
Haloperidol 4 mg BID 88.5 (1.0)
Change from Baseline -10.7 (1.6)
-16.2 (1.7)*† -14.9 (1.7) -15.4 (1.6)*89.0 (0.9)
Salmon
Dr. Helm,
I agree that the change in negative scores on the PANSS is likely not clinically significant.
The following are PANSS data from the FDA reviews of the latest antipsychotic approved.
Baseline Mean (SE)
PBO 89.0 (0.9)
Asenapine 5 mg BID 88.9 (1.0)
Asenapine 10 mg BID 89.4 (1.0)
Haloperidol 4 mg BID 88.5 (1.0)
Change from Baseline
PBO -10.7 (1.6)
Asenapine 5 mg BID -16.2 (1.7)
Asenapine 10 mg BID -14.9 (1.7) NS
Haloperidol 4 mg BID -15.4 (1.6)
Approx Score at Endpoint
PBO 78.3
Asenapine 5 mg BID 72.7
Asenapine 10 mg BID 74.5
Haloperidol 4 mg BID 73.5
As you can see the average difference in all symptoms between drug and placebo is really not very large, i.e. on the order of around 5 points.
If most of the response is due to effects on positive symptoms then the difference in effects on negative symptoms between asenapine and haldol is at most likely 1 point or so which is really clinically insignificant when considering that the symptom score for all symptoms is still in the 70 point range. Plus as indicated by the standard error in the measurements this amount of difference may be purely due to chance.
Because of this new scales are being developed to show effects on negative symptoms but in truth this just cuts things finer. You can think of this in a simplistic way as just multiplying the PANSS by 10. So instead of 90 at baseline and 75 after treatement it’s instead 900 and 750 respectively. Consequently you can more easily show that a 1 point difference in negative scores is statistically significant as it’s now a 10 point difference. But it’s really just due to using a finer measure and likely has little to no clinical implications.
In my experience when you look at the raw data, there’s often little difference between drug and placebo and sometimes the difference is due to a few individuals doing extremely well on drug whereas almost everyone else you can’t tell the difference.
I believe the drugs work and have seen dramatic improvements in the clinic in some individuals but we shouldn’t allow ourselves to be fooled that good response in a few individuals means that the vast majority of people respond any better to drug than placebo.
Maybe if all raw data had to be released more intelligent usage might be possible.
Salmon
M Helm, MD
Salmon, thanks for the data and insight. Your reporting of the newest med is very close to my recollection of data on the other meds.
Maybe you are right that if the raw data were released there would be more rational use. However, I’m afraid that even if it were more available, there would be few clinicians who would take the time to understand the information. The marketing messages are more easily understand, and (because we always want to believe that newer is better) more likely to be believed.
My take-home from the above is that asenepine is about the same as haloperidol. There should be a far greater knowledge and experience base of the complexities of managing haldol than asenapine, but I’d bet that (excluding the VA system) it won’t be long before there is more money spent on asenapine than on haloperidol.