Saturday, August 27, 2011

Drug Trials in 'At Risk' Youth

Is it ethical to medicate healthy teenagers "at risk" of developing psychosis to prevent a symptom that may not occur? One such clinical trial in Australia was recently stopped before it could even begin:
Drug trial scrapped amid outcry

Jill Stark
August 21, 2011

FORMER Australian of the Year Patrick McGorry has aborted a controversial trial of antipsychotic drugs on children as young as 15 who are "at risk" of psychosis, amid complaints the study was unethical.

The Sunday Age can reveal 13 local and international experts lodged a formal complaint calling for the trial not to go ahead due to concerns children who had not yet been diagnosed with a psychotic illness would be unnecessarily given drugs with potentially dangerous side effects.

Quetiapine, sold as Seroquel, has been linked to weight gain and its manufacturer AstraZeneca, which was to fund the trial, last month paid $US647 million ($A623 million) to settle a lawsuit in the US, alleging there was insufficient warning the drug may cause diabetes.
Dr. McGorry works at Orygen Youth Health in Parkville (near Melbourne) and is a proponent of early interventions for treating mental illness and substance abuse (McGorry et al., 2011). In discussing psychotic disorders, these authors say:
The importance of timely treatment initiation has been further underscored by new data from the Treatment and Intervention in Psychosis (TIPS) project showing that early treatment had positive effects on clinical and functional status at 2-year and 5-year follow-up in first episode psychosis. These studies showed that reducing the duration of untreated psychosis has longer-term effects on the course of negative symptoms, depressive symptoms, cognitive symptoms and social functioning, suggesting the possibility of secondary prevention of these pathologies in first-episode schizophrenia.
But how about prevention, as opposed to early intervention? Are researchers and clinicians able to predict (with reasonable accuracy) who will develop schizophrenia? The scrapped clinical trial intended to see whether quetiapine would decrease or delay the risk to 15-40 yr old participants who showed "early signs" of developing a psychotic disorder. What are these early signs and prodromal symptoms (Mechelli et al., 2011)?
...a gradual deterioration of global and social functioning and the emergence of attenuated psychotic symptoms. However, not all people with these features progress to develop a full-blown psychotic disorder; 20% to 50% develop psychosis, usually within 24 months, but the remainder do not. Individuals first seen with this clinical syndrome are, thus, said to be at ultra-high risk (UHR) for psychosis.
So anywhere from 50% to 80% of those showing prodromal symptoms and labeled ultra-high risk do not develop a psychotic disorder such as schizophrenia. Can neuroimaging improve this crude level of prediction? Ultimately, disordered thinking, delusions, and hallucinations arise from the brain -- right? -- so we should be able to see some abnormality on an MRI scan. A multi-site study enrolled 182 individuals at UHR for psychosis and 167 healthy controls. Voxel-based morphometry was used to quantify whole brain gray matter volumes, as well as 3 specific regions of interest (ROIs): the left parahippocampal gyrus in the medial temporal lobe (important for memory), the right inferior frontal gyrus (important for attention and cognitive control), and the left superior temporal gyrus (which may be implicated in auditory verbal hallucinations). Two years later, 48 UHR participants (26%) developed psychosis and the others did not.

Compared to controls, all subjects in the UHR group showed gray matter reductions in medial frontal regions, so this result was not predictive of whether full-blown psychosis would occur. Within the UHR group, however, a tiny region in the left anterior parahippocampal gyrus (6 voxels) differed in the UHR who later developed psychosis and those who did not. There were no volume reductions in the other two ROIs.

Figure 2 (adapted from Mechelli et al., 2011). Differences between ultra-high-risk (UHR) individuals who did (UHR-T) and did not (UHR-NT) develop psychosis. The UHR-T individuals had less gray matter volume than did the UHR-NT individuals in the left parahippocampal gyrus, bordering the uncus (MNI [Montreal Neurological Institute] coordinates x, y, and z: –21, 6, and –27, respectively). For visualization purposes, effects are displayed at P < .05 uncorrected.

So basically, only 6 voxels in the entire brain were capable of predicting whether or not a patient at ultra-high risk for psychosis will indeed be one of the 26% to progress to a clinically significant psychotic disorder. To examine the actual diagnostic accuracy, the authors then took gray matter volumes from the peak voxel, performed cross-validation analyses using a predictive linear model, and determined that the average predictive accuracy was only 62% (sensitivity = 61% and specificity = 65%). This means the single voxel measure incorrectly predicted that 39% of healthy UHR would become psychotic, while it missed a diagnosis in 35% who would later develop psychosis.

On the basis of these results would you recommend MRI scans of the left parahippocampal gyrus to refine the cohort given Seroquel to reduce or delay the risk of psychosis? I would say no, especially not if you're Australian (Dazzan et al., 2011). Another paper by many of the same authors reported that of 102 UHR Australians, 28 converted to a psychotic disorder, and these individuals showed volume reductions in frontal [and other] regions, relative to the UHR subgroup who remained healthy (Dazzan et al., 2011). Left parahippocampal gyrus was nowhere to be found. In fact, none of the 3 ROIs from Mechelli et al. were selected as ROIs by Dazzan et al. Furthermore, neither of these papers cited the other, despite the fact that they shared 8 authors in common.

In my view, the results thus far seem disappointing to those looking for the neuroanatomical correlates of psychiatric disorders. What does this mean for future structural MRI studies searching for changes that will predict the onset of psychosis?


Dazzan P, Soulsby B, Mechelli A, Wood SJ, Velakoulis D, Phillips LJ, Yung AR, Chitnis X, Lin A, Murray RM, McGorry PD, McGuire PK, Pantelis C. (2011). Volumetric Abnormalities Predating the Onset of Schizophrenia and Affective Psychoses: An MRI Study in Subjects at Ultrahigh Risk of Psychosis. Schizophr Bull. Apr 25. [Epub ahead of print]

McGorry PD, Purcell R, Goldstone S, Amminger GP. (2011). Age of onset and timing of treatment for mental and substance use disorders: implications for preventive intervention strategies and models of care. Curr Opin Psychiatry 24:301-6.

Mechelli, A., Riecher-Rossler, A., Meisenzahl, E., Tognin, S., Wood, S., Borgwardt, S., Koutsouleris, N., Yung, A., Stone, J., Phillips, L., McGorry, P., Valli, I., Velakoulis, D., Woolley, J., Pantelis, C., & McGuire, P. (2011). Neuroanatomical Abnormalities That Predate the Onset of Psychosis: A Multicenter Study. Archives of General Psychiatry, 68 (5), 489-495 DOI: 10.1001/archgenpsychiatry.2011.42

Subscribe to Post Comments [Atom]


At August 29, 2011 12:21 PM, Anonymous Anonymous said...

This is a pretty limited summary of the much larger field of risk prediction for psychosis. There are a number of papers showing that there are different developmental trajectories in high risk individuals who go on to develop psychosis- there have been a number of structural MRI studies in addition to diffusion tensor imaging.

An ONGOING goal of the field is to use these findings to be able to predict who will and will not transition to psychosis. This is not trivial- part of the difficulty is that in almost all studies involving this population, the patients are receiving some kind of treatments- family therapy, individual therapy, often they are on whatever combination of antidepressants/anxiolytics/etc that their own Dr. recommends (because to deprive them of those things would be unethical). And it can be hard to know whether the act of identifying and treating at risk individuals changes the proportion of people who become psychotic- it is an unavoidable confound. The other difficulty is that often follow up periods are for 2-3 years, and that means there are people in the "unconverted" groups who may ultimately also get psychosis after the study is complete. The only real way around this is with very very large sample sizes, such that the sample size of individuals who convert becomes big enough to make conclusions about. To this end, the field has now moved towards very large multi-site consortiums, most of which are still gathering data.

It is premature to damn this field based on a limited reading of the literature... no one doing this work thinks it is "done", and no one would use the current MRI findings to target treatment approaches... but that doesn't mean that we won't get there eventually. Have patience!

At August 29, 2011 1:36 PM, Anonymous John M. Nardo MD said...

Thanks! An excellent summary. Running on his charisma and enthusiasm, McGorry is trying to outrun his science...

At August 29, 2011 7:37 PM, Blogger The Neurocritic said...

Anonymous - Thanks for your comment. The post isn't meant to be a comprehensive overview of MRI studies of risk prediction for psychosis. I just chose the two most recent papers that included McGorry as an author. I am a neuroscientist and do believe there must be a brain basis for psychotic disorders. But I'm not an expert in the field, so I was struck by the contradictory findings in these particular papers.

I also realize that longitudinal studies of this magnitude are difficult and expensive to conduct, and that patient selection, diagnosis, and retention are very challenging issues. In fact, one of the differences between the two papers (other than multi-site vs. Australia only) was the ultimate diagnosis the UHR group would convert to: "psychosis" in Mechelli et al. (n=48) vs. schizophrenia (n=19), bipolar disorder with psychotic features (n=2), or major depression with psychotic features (n=5) with the latter two collapsed to form a small group with affective psychosis in Dazzan et al. Which brain changes should you be looking for if you can't predict the diagnosis in advance? The ones seen in generic psychosis? Or the ones in schizophrenia and the various affective psychoses? All of them?

Any suggestions for further reading would be helpful...

John M. Nardo MD - Yes, it seems that taking Seroquel if you don't really need it isn't such a great idea!

At August 30, 2011 11:01 PM, Anonymous Anonymous said...

how can I get in touch with you via email?

At August 31, 2011 8:27 AM, Blogger The Neurocritic said...

What is your e-mail address? I won't publish it...

At September 02, 2011 4:40 PM, Anonymous Anonymous said...

Anonymous #1 here... I can give you some other reading ideas if you'd like I just have to round up a list. Sorry my comment turned out to be so long! (that second anonymous wasn't me).

You're totally right, the issue with diagnosis is really tough in this case. It's hard enough in full on adult patients, when you have things like schizophrenia, schizoaffective disorder, schizophreniform disorder, bipolar with psychotic features, depression with psychotic features, etc. And in an adolescent who is newly diagnosed, it's that much harder. But, it seems like there is more overlap, genetically and in terms of mechanisms and phenotypes, between these things than previously thought- particularly between schizophrenia spectrum disorders and bipolar with psychotic features. So, in some ways, in line with the new spectrum based direction of the DSM, "psychosis" isn't a bad way to go, even though it sounds kind of vague initially.

At September 03, 2011 12:16 AM, Blogger Lili Marlene said...

The idea that in real cases there appears to be a lot of overlap in symptoms between mental illnesses like schizophrenia and depression and bipolar is not a new idea. Years ago I read a book about schizophrenia which had this in it.

At May 30, 2014 4:26 AM, Blogger Unknown said...

And it can be hard to know whether the act of identifying and treating at risk individuals changes the proportion of people who become psychotic- it is an unavoidable confound.


Post a Comment

<< Home

eXTReMe Tracker