Do ADHD Drugs Work Longterm?

“Illness is the doctor to whom we pay most heed; to kindness, to knowledge, we make promise only; pain we obey.” ― Marcel Proust

An essay on ADHD in the New York Times launched an interesting Twitter exchange with Steve Silberman and a medical blogger PalMD on how well we understand psychiatric disorders and treatment.

In the article, Dr. Sroufe concludes that since there is no evidence for longterm use of ADHD medication, their use should be abandoned. He is right that the evidence of efficacy is all short term. Over the long term, no benefit has been shown. Of course almost no one dealing with the issue on a day to day basis would agree. Parents, teachers and physicians all agree that these medications have a use to improve the lives of these children. Count me among those who believe it is highly probable that treatment over the course of months and years has utility, but is hard to prove.

As a problem in decision making, this is a good example of the difference between believing and knowing.

There is a difference between the practice of science and an absolutist approach to truth. In decision making, we must be practical. As Williams James said, “Truth is what works.” He believed that science was a pragmatic search for useful models of the world, including mind. Those that look for abstract, absolute truth in clinical research will be confused, misguided and as often as not, wrong in their decisions. Truth is something that happens to a belief over time as evidence is accumulated, not something that is established by a single positive experiment.

Belief in the usefulness of therapy in medicine follows this model of accumulation of belief. The complexity and variability of human behavior demands a skeptical approach to evidence and a sifting through to discover what works.

Clinical trials for drugs to affect behavior are generally relatively small, short experiments that measure a change from baseline in some clinically meaningful variable. These trials are clinical pharmacology studies in the classic sense, studies in patients (clinical) of drug effect (pharmacology). No one is expecting cure or even modification of the disease. The benefit is short term symptom relief, so the trial examines short term symptom relief. In the case of a pain reliever, we ask whether patient’s self reports of pain are decreased by therapy compared to before therapy. In ADHD, we ask whether a group of target behaviors is changed by treatment compared to baseline.

This approach of measuring change from baseline has a host of pitfalls that limit the generalizability of clinical trials to real life medicine. First, baseline measures are subject to large amounts of bias. One of the worst sources of bias in these trials is the patient and physician’s joint desire to have the patient meet the severity required to be enrolled. The investigator is under pressure to contribute patients to the trial. The patient hopes to gain access to some new therapy, either during the trial or during some subsequent opportunity. Both of these factors pressure patients to maximize the severity of their complaint at baseline. How do you get into a trial? Exaggerate your problem! Even without conscious or unconscious bias from patients, any trial will enroll patients that happen to be worse than their average state. When measured repeatedly over time, the scores will tend to drop- a classic regression to the mean. If you select more severe outliers, they will tend to look more average over time.

Second, diseases are not stable over time. Without any intervention, measures of a disease will be most highly correlated when measured with a short duration between assessments. The longer you wait to measure again, the lower the correlation. Measuring a drug effect in a controlled trial accurately depends on a high level of correlation. All else being equal, the longer one treats, the harder it will be to measure the effect of the drug. This is the major pitfall of depression trials. Episodes are usually limited in duration, so most patients will get better over time without treatment.

So perhaps its not surprising that its very hard to measure the effect of ADHD drugs after months or years in chronic therapy trials. These kids get better over time both from regression to the mean and the natural history of the disease.

Another important issue in ADHD research is that these drugs have effects in healthy volunteers. As Dr. Sroufe points out, amphetamines help college students study for exams- no diagnosis of ADHD needed. This makes it easier to do pharmacology studies, but means that diagnosis in those studies doesn’t really matter- the pharmacology is largely independent of any real pathological state. One could never study a cancer drug in some one without cancer, but this is not true of a cognitive enhancing drug. Its probably most likely that the kids with ADHD don’t have a single pathophysiology, but rather a combination of being at one end of a normal spectrum of behavior plus stress or lack of coping mechanisms that create problems for them in the school environment where those behaviors are disruptive to their learning and that of others. The pharmacology of stimulants helps then all- after all it helps even neurotypic college students and computer programmers.

Treatment response does not confirm diagnosis in ADHD as it does in some other neurological diseases like Parkinson’s Disease. While we’d like to call ADHD a disease or at least abnormal brain state, we have no routine way of assessing the current state of a child’s brain. We have even less ability to predict the state of the brain in the future. Thus diagnosis, in the real meaning of the word- “dia” to separate and “gnosis” to know, is something we can’t do. We don’t know how to separate these kids from normal or into any useful categories. And we have no way of describing prognosis- predicting their course. So a trial that enrolls children on the basis of a behavior at a moment in time and tries to examine the effects of an intervention over the long term is probably doomed to failure. Many of those enrolled won’t need the intervention over time. Many of those who don’t get the intervention will seek other treatment methods over time.

With all of these methodological problems, we can’t accept lack of positive trials to be proof that drugs are ineffective long term. We can’t even prove that powerful opioid pain relievers have longterm efficacy. In fact, it was not too long ago that we struggled with a lack of evidence that opioids were effective even over time periods as short as 12 weeks.

Our short term data in ADHD provides convincing evidence of the symptomatic effects of treatment. Instead of abandoning their use, we should be looking at better ways to collect long term data and test which long term treatment algorithms lead to the best outcomes. And we should be using our powerful tools to look at brain function to understand both the spectrum of ADHD behaviors and the actions of drugs in specific brain regions.

Why Enrichment Designs Don’t Work in Clinical Trials

Last week I was discussing a clinical trial design with colleagues. This particular trial used an enrichment design. A few years ago I did some simulation work to show that you can’t pick patients to enroll in a clinical trial in order to improve the results.

People are probabilistic too.

The idea of and enrichment design is to winnow the overall patient group down to those individuals who are likely to respond to therapy. One way is to give all of the candidates a placebo and eliminate placebo responders. Another strategy is to give a test dose of drug and keep only those who respond. Either way, the patients that pass the screening test get to go on to a double blind test of active drug versus placebo.

Sounds like a great idea, but it doesn’t really work most of the time in practice. While this idea of screening out patients, it turns out that it mostly just excludes patients who are varying in their complaints over time. You can’t really  tell who are going to be better patients during the screening test. It turns out that most patients look different at one time point compared to any other.

The mistake that we make is in thinking that people can be categorized by simple inspection. We think of patients as responders or non-responders, an intrinsic characteristic they have or don’t have. Trying to screen out patients we don’t want falls into the trap of thinking that a single set of tests can successfully discriminate between classes.

The way I think of it is that we need relatively large clinical trials to prove the value of a modestly effective drug. So it seems odd to think that one could easily categorize patients themselves when tested. You can see this by looking at how well a test dose of a drug looking for drug responders would be able to enrich a patient population. Variability over time makes this impossible.

Let’s walk through an example. An imaginary trial of a drug to treat migraine attacks.

Lets say we know the truth and this candidate is in reality a pretty good treatment for a migraine attack. But the patient varies in headache severity and responsiveness to treatment.

Some headaches are mild and will resolve without treatment. That mild attack will act no differently whether the active drug or placebo was administered. Some headaches are very bad and even a really effective drug might not touch that kind of headache. So again the attack will be the same whether placebo or treatment is given.

And what about the headaches that are in between and could respond? Well if a drug worked half the time, then out of every two of those attacks, the active drug would show an effect where the placebo did not. The other half the time, it would look just like placebo again.

Add up these cases, there are four of them. For only one atttack did the active drug work where the placebo would fail. One out of 4 times, a 25% overall response rate. All just because in the same patient the headache and its response to drug changes. So if I did a test treatment to see if I had a responder, I would eliminate half of the responders because either they had a headache that was the the one too severe to respond or the one that happened not to respond that time.

Of course you’d eliminate some of the non-repsonders. But we know that even non-responders may have 1 in 4 headaches that are mild enough that they don’t need the treatment anyway. So you eliminate 75% of the non-responders with a test dose which is better than the 50% of responders that were eliminated. You’ve done better. How much better depends on the ratio of responders to non-responders in the population, a ratio that is completely unknown.

What’s nice is that while you can see the logic by reading the story I’ve told, a mental simulation, one can create an explicit mathematical model of the clinical trial and simulate running the trial hundreds of times. It turns out that there very few conditions where this kind of enrichment really works. I turns out its simpler and just as informative to see whether or not the drug is effective in the overall population without trying to prejudge who is a responder or not with a test dose.

The irony? This is exactly the opposite of clinical practice. In the real clinic, each patient is their own individual clinical trial, an “N of 1” as we say. N is the symbol for the number in a population. An individual is a population of one. N of 1. We treat the patient and over time judge whether or not they respond in a personal clinical trial. Not to see whether the drug works but whether the patient is a responder.  If they don’t, therapy is adjusted or changed. But in our migraine example, multiple headaches of various intensity would have to be treated so see the benefit.

Perhaps variability across a population is easily grasped. People are tall or short, have dark or light hair color. Variability within an individual over time is perhaps more subtle but just as important for over time.

Neurology Notes

Not Practicing

It looks like I’m not alone in Maryland as a non-practicing physician:

State lacks practicing physicians — “While the state has about 25,000 licensed physicians, the second-highest rate per capita of any state, nearly 40 percent are engaged in teaching, research and administrative duties, according to the study, and some of the rest spend part of their time in such nonclinical work.”

One the insights that I had many years ago was that in practice, one had no way to leverage one’s activities. Economically a physician is valued for their direct time in providing care. And procedural activity, like surgery or angioplasty, is valued per unit time way above examining and diagnosing. In research and business, there are value multipliers as others participate in the enterprise. The return on time is potentially much higher for a phyisician outside of practice.

Low FDA Approval Rates

The rate of new drug approvals for the pharmaceutical industry is dropping. One reason may be FDA delay:

Eye on FDA: Impact of Approvable Letters 2007 – Part 1: “However, it is important to think about the implications of so many approvable letters.  They represent more than an inconvenience to companies and to patients who are awaiting therapies with serious implications for both, as well as for investors in companies, particularly smaller companies that do not yet have a product on the market.  ”

The drug is safe and effective, but is not yet allowed to market. This in between status for a drug can drag on for years, destroying companies.

Neurology Notes

Worth the Subscription

I subscribe to the online edition of The Lancet Neurology. The 2007 Round-Up was worth the subscription price itself. I need to keep up very broadly and these are the kind of overviews that help me the most.

Also in The Lancet Neurology:

The Lancet Neurology: “We used a unique response-conditional crossover design to provide rescue treatment if needed, and patients who showed an improvement in inflammatory neuropathy cause and treatment (INCAT) disability score during treatment were re-randomised into a 24-week extension phase.

I’m reading this both to gain some insight into immunologic treatment successes, but I’m trying to figure out how useful the trial’s design might be. I’ve been skeptical of enrichment designs in the past, but this may be a more clinically relevant way to study the reality of clinical responses.


I did my MD, PhD training as well as my medical internship at Emory School of Medicine in Atlanta. I’m sorry, but not surprised that it may close.

A Safety-Net Hospital Falls Into Financial Crisis – New York Times: “Once admired for its skill in treating a population afflicted by both social and physical ills, Grady, a teaching hospital, now faces the prospect of losing its accreditation. Only short-term financial transfusions have kept it from closing its doors.”

When I trained, Grady was a resident’s hospital. Care was provided by residents under the direction of Chief Residents who had completed training and were often specialty fellows continuing training and a group of incredibly knowledgeable teachers who ran the services and provided checks on care through morning rounds. There were teaching attending physicians, but their function was to teach and supervise, not provide care.

After I left, medical reimbursement rules changed so that the attendings had to provide care and sign off everything if the school was going to be able to bill at all. Having to switch from resident care to medical faculty care kicked a big prop out from under academic medicine and was one of the forces that eventually led me to my career in industry.

Neurology Notes

Video: Trepanation

Retrospectacle: A Neuroscience Blog: “Trepanation is a procedure where a hole is drilled into the skull, exposing the dura mater and brain for either medical (releif of pressure) or mystical (supposed heightened consciousness)purposes. ”

Pharma Marketing

For reasons I don’t really understand, this report is getting lots of attention in the media and pharma-related weblogs. These are large businesses with shareholders expecting return on investment. They market where they believe the return will be found. The same is true of R&D and the relative rate of spend between the two.

Big Pharma Spends More On Advertising Than Research And Development, Study Finds: “A new study by two York University researchers estimates the U.S. pharmaceutical industry spends almost twice as much on promotion as it does on research and development, contrary to the industry’s claim.”

Up and Coming: Metabotropic Glutamate Receptor Modulation

It seems that since the September announcement by Lilly showing efficacy of a metabotropic glutamate receptor agonist in Schizophrenia, there’s more activity in the area than I’ve ever seen before:

Pfizer’s in:

Pfizer Inc. and Taisho Pharmaceutical Finalize Deal on Schizophrenia Drug; to Pay Initial $22M – News, Search Jobs, Events: “TS-032 is a novel mGluR (metabotropic glutamate receptor) agonist that may offer a new treatment option for central nervous system disorders. Although the characteristics of mGluR are still only partly understood, mGluR is believed to play a role in the transmission of glutamate and other substances in the brain. Abnormalities in the neurotransmission through mGluR may be one cause for symptoms related to schizophrenia as well as other CNS disorders. Data show that mGluR agonists, such as TS-032, offer potential as new treatments for schizophrenia.”

As is Merck:

Addex Pharmaceuticals : 3 January 2008 ADX63365: “Allosteric modulation company Addex Pharmaceuticals (SWX:ADXN) announced today that it has entered an exclusive worldwide license agreement with Merck & Co., Inc. (“Merck”) to develop ADX63365, an orally available drug candidate for the potential treatment of schizophrenia and other undisclosed indications. Allosteric modulators are an emerging new class of therapeutic agents. ADX63365, currently in preclinical development, is a positive allosteric modulator (PAM) that targets the metabotropic glutamate receptor 5 (mGluR5), which is believed to be important as a target for the treatment of schizophrenia and other conditions. The deal also includes mGluR5 PAM backup compounds discovered by Addex.”

Sol Snyder on Seeking God in the Brain

It seems to be Guilford Pharmaceuticals day here. Sol Snyder, scientific founder of Guilford has this wonderful piece in the New England Journal:

NEJM — Seeking God in the Brain — Efforts to Localize Higher Brain Functions: “In seeking a general relationship between religious states, poetry, and music, Trimble ascribes all three to the right, nondominant side of the brain. He assumes that integration of the activity of the right-sided emotional brain with that of the left-sided analytic brain gives rise to the greatest intellectual achievements in the arts. I suspect that major advances in science, too, are the product of more than pure reason — in the finest scientists I have encountered, I have always detected a notable creative, artistic flair. ”

Can we be better scientists by practicing art? The specificity principle of general adaptation would suggest that scientists would be best practicing science like art.

A randomized clinical trial of coenzyme Q10 and GPI-1485 in early Parkinson disease — The NINDS NET-PD Investigators 68 (1): 20 — Neurology

On the subject of Guilford Pharmaceutical drugs that I’ve worked on, the second NIND NET-PD trial has been published, a futility design that included GPI-1485.

A randomized clinical trial of coenzyme Q10 and GPI-1485 in early Parkinson disease — The NINDS NET-PD Investigators 68 (1): 20 — Neurology: “Coenzyme Q10 and GPI-1485 may warrant further study in Parkinson disease, although the data are inconsistent. ”

This trial was independently conducted by NINDS. We supplied the drug. Their results in untreated patients were consistent with our own 2 year long study of patients who were already on dopamine agonist monotherapy. There are important lessons to be learned from these studies about the challenges of designing studies to test disease modifying drugs in Parkinson’s Disease and other neurodegenerative diseases.

Aetna to End Payment for a Drug in Colonoscopies – New York Times

Of interest:

Aetna to End Payment for a Drug in Colonoscopies – New York Times: “Aetna, one of the nation’s largest private health plan managers, is the latest insurer to clamp down on the use of a powerful anesthetic during an increasingly common form of colon cancer screening.”

This insurer’s reaction to the use of anesthesiologists to administer propofol for colonoscopy is of interest to me because of my work with AQUAVAN over the years. At Guilford Pharmaceuticals (and then MGI Pharma after their acquisiton of Guilford) I directed the first in man study of AQUAVAN and worked on trials on and off through the recent NDA filing. AQUAVAN, as a prodrug of propofol, is not expected to require the same FDA mandate for administration by an anesthesiolgist that propofol does.

AQUAVAN is a prodrug of propofol, extending its action enough to make it a better sedative and a less powerful anesthetic. It’s an interesting, perhaps unique example of how making a drug worse for one use makes it more suitable for another.

This Times article doesn’t mention the alternatives like AQUAVAN or midazolam, but with pressures like this on physicians from insurers and pressures from patients for comfort during procedures, I expect that AQUAVAN will find its niche as a sedative for procedures. Eisai is now in the process of acquiring MGI, so AQUAVAN’s future now rests with them.

CRO Industry Outlook

Investor’s Business Daily: Upturn In Biotech Spending Drives Covance’s Growth: “Growth in the CRO sector is driven by biopharmaceutical research, development spending and a robust biotech funding environment, which hit $20.1 billion last year.”

As large pharmaceutical companies struggle with their cost structures they turn increasingly to outsourcing. Development stage biotechs often choose not to build their own infrastructure from the start to avoid the fixed costs. It’s driving CRO growth, where I currently work.

Mind Hacks: Personalised drugs

Mind Hacks: Personalised drugs:

The idea of genetically testing people for drug suitability is causing them [Drug Companies] a bit of a headache at the moment, as they’re desperately trying to think of ways to make money out of it.

Pharmacogenomics has to be one of the most misunderstood areas of drug development today. We’re used to the idea by now that our genes do not determine who we are. Genetic inheritance puts us at risk for some diseases more than others, makes us more or less likely to excel at certain mental or physical tasks, influences our adult height or weight. But strong effects of single genes are rare. Instead there’s a complex interacting system of multiple genes and environmental effects that, based on what we now know about complex systems, will not act deterministically, but rather affect the probability of future events.

It seems clear that, unless there is a strong single gene effect on something like proteins involved in drug metabolism and clearance, genes will have an uncertain influence on response to drug.

In the end, knowing some one’s genetic background, like knowing their particular symptom complex, will inform the physician about where to start therapy and the chances of success. But there will never be a way of “knowing” if that means a high degree of confidence in knowing the outcome of therapy.