An Expert Interview with William T. Carpenter, Jr., MD
Editor's Note:
The single-disease paradigm does not fully explain the pathophysiology of schizophrenia, nor does it allow for the most comprehensive treatment of patients with this condition. On behalf of Medscape, Randall White, MD, speaks with William T. Carpenter Jr, MD , Professor of Psychiatry and Pharmacology, University of Maryland School of Medicine, Baltimore Maryland, about why the etiology and treatment of cognitive and negative symptoms must be considered as individual domains separate from psychosis.
Medscape: Is it correct to say that, in your view, the single-disease paradigm of schizophrenia is flawed and has brought us to a nosological impasse?
William T. Carpenter Jr, MD: Yes, I sort of think that, but I'd put it a little bit differently. I would say that the single-disease entity has not been confirmed, and there are many reasons to have doubts about this idea and to expect that there is more than 1 etiologically based disease within the syndrome. The single-disease paradigm presumes that the different aspects of pathology are coming from the same latent structure and, therefore, would be similar both in etiology and in treatment response. The idea of a single disease entity has been a very weak heuristic. It has led to a focus on psychosis and a relative neglect of cognition impairments and negative symptoms. We do know that for treatment considerations, those are remarkably different aspects of pathology.
Medscape: You advocate looking at schizophrenia as having several domains of psychopathology. What are those?
Dr. Carpenter: This idea came from the international pilot study conducted years ago when I was working with John Strauss and John Bartko. We found that the measures of psychosis did not predict either how much of the other aspects of illness a person would have (like negative symptoms and interpersonal pathology) or future [outcome]. Each of these domains had its own history, its own nature of manifestation, and its own future. You could predict within, but not across, a domain.
The concept has been refined since then and includes cognition impairments, the type of impairments that you pick up on neuropsychological test assessment. It could also be that disorganization of thought and behavior is different from the reality-distortion symptoms in an important way. Rather than studying schizophrenia as a single disease -- expecting that 1 aspect of the condition would generalize to the others -- each of these domains should be studied in its own right in terms of both cause and treatment.
Medscape: What further research is needed to test the paradigm of multiple domains? And, are these domains independent?
Dr. Carpenter: As a construct, it's best to act as though they are independent and leave, for some future resolution, [the question of] why they co-occur in the illness syndrome. I think we'll be much more likely to understand etiology, pathophysiology, and make treatment discoveries if we make each domain its own target.
With that said, I doubt if these domains are truly independent. It could be that, for example, 1 disease process causes both psychosis and negative symptoms, while another disease process causes psychosis without negative symptoms.
Medscape: You have encouraged a distinction between negative symptoms and deficit symptoms. Can you explain?
Dr. Carpenter: Negative symptoms, probably best characterized as restricted affect, alogia, diminished drive, anhedonia, and avolition, can occur for a number of reasons. When first discussed, negative symptoms were meant to be direct manifestations of the schizophrenia itself. But, as the rating scales that captured them came into use, it turned out that people were rating negative symptoms regardless of whether they were caused by schizophrenia or not. For example, if a person is depressed, he or she will be anhedonic, and when over the depression, will be less anhedonic or may not be anhedonic at all. The state-dependent negative symptom is very different from the concepts developed by Kraepelin and Bleuler, who described a reduced capacity for these ordinary behaviors as a trait phenomenon in schizophrenia.
As another example, someone who is paranoid and socially withdrawn is rated as having negative symptoms. As she becomes less paranoid and, therefore, more socially engaged, it looks like her negative symptoms have improved, but it's all secondary to the [change in] paranoia. Similarly, a drug like haloperidol, particularly at higher doses, will induce negative symptoms. We introduced the term "deficit pathology" to refer to negative symptoms that are direct manifestations of the schizophrenia, not negative symptom ratings that can occur for a variety of reasons in people with schizophrenia but are not caused by the schizophrenia itself.
Medscape: Are there studies that permit an understanding of the long-term stability and prognosis of the deficit syndrome in schizophrenia?
Dr. Carpenter: We need to learn a lot more, but there are some things that we do know already. An aspect of this is a bit tautological. By definition, primary negative symptoms have to be present as trait pathology to meet the criteria. And, if the symptoms are present as trait, they're likely to be long-lasting into the future. So, the presence of primary negative symptoms will predict, in the long term, the presence of primary negative symptoms. If in fact there are rapidly fluctuating primary negative symptoms, we don't have a good way to detect them.
Medscape: Researchers such as Nancy Andreasen and Timothy Crow have suggested neurological features associated with deficit symptoms, like ventricular enlargement. Do you believe these findings have prognostic or therapeutic significance?
Dr. Carpenter: Well, in Crow's proposition, the difference between type I and type II schizophrenia would essentially be the presence or absence of irreversible negative symptoms. In that regard, it would be similar to how we think about deficit and nondeficit schizophrenia. I think his hypothesis that this was based on cell loss and would not be responsive to antipsychotic drug treatment did not prove to be correct.
Volume loss, particularly in the hippocampus, and increased ventricular volume are fairly widespread in schizophrenia. I don't believe they are restricted to the deficit subtype. In fact, in our studies, when we've isolated the deficit subtype, we've tended not to see as much volume loss as in the nondeficit form of schizophrenia. We've actually seen a tendency toward enlargement in other structures in the deficit form. It seems, therefore, that the neuropathology of schizophrenia differs between those with vs those without the deficit features.
Medscape: Enlargement in which structures?
Dr. Carpenter: In 1 study, there was enlargement in lobules 8, 9, and 10 of the cerebellum. In both normal [subjects] and nondeficit schizophrenics, the larger that structure, the better the performance on cognitive tests, and the nondeficits were not different in size from the normal controls. The deficit subgroup had larger lobules 8, 9, and 10, and, in this group, the larger that structure, the worse the cognitive performance -- as if there was a neuropathologic process that simultaneously made it larger and interfered with function.
Robert Buchanan's magnetic resonance imaging [MRI] studies tended to show greater volume losses in the nondeficit cases than in the deficit cases, and greater enlargement of the head of the caudate in deficit cases than in nondeficit cases.
Medscape: What do we know about long-term stability and prognostic implications of the cognitive dysfunction of schizophrenia?
Dr. Carpenter: Most, and perhaps even all, patients with schizophrenia have impaired cognition. It may be that some subgroups are spared but, in general, it looks like schizophrenia is associated with a reduction in IQ in the neighborhood of 5 to 10 points. It starts relatively early in life and is probably in place by the time people begin to manifest psychosis.
These impairments also appear to mark people at high risk for schizophrenia. It is possible that they begin, in some way, from birth, but the impairments certainly appear well in advance of psychosis and tend to be a stable trait throughout the course of the illness. You might conceptualize schizophrenia as an illness that begins not in late adolescence or early adulthood -- only the psychosis domain becomes manifest during that time period, while cognitive pathology manifests much earlier.
We think of cognitive impairment as trait pathology and as having an uncertain relationship with negative symptoms. Our studies and some others suggest that the deficit subgroup will have more impairment in tests that tap into frontal and parietal functions. But, basically, we just don't know the relationship between cognition and negative symptoms, largely because most studies have used measures that don't separate primary and secondary negative symptoms. There is a very low relationship between the cognition impairments and psychosis.
Medscape: What are the therapeutic implications of your domains-of-pathology model?
Dr. Carpenter: If you attempt to identify primary negative symptoms and look at the treatment studies to date, there's very little documentation of drug treatments having efficacy for primary negative symptoms. The negative symptom improvements that are routinely reported include secondary negative symptoms and may actually be restricted to secondary negative symptoms. The studies that have attempted to identify primary negative symptoms have not been able to document therapeutic efficacy of current pharmacotherapy.
Incidentally, any advantage in secondary negative symptoms is a true advantage, but theoretically it makes a huge difference as to whether or not you try to refine your antipsychotic drugs to treat negative symptoms of schizophrenia or whether it's an unmet therapeutic need that requires an entirely different discovery approach. The same can be said for cognition impairments. A number of drugs will look better on cognition and negative symptoms than the comparator haloperidol, but that may relate principally to the negative effects of haloperidol on those aspects of illness rather than the true efficacy for either cognition or negative symptoms of the newer treatments. So, the construct has helped us critique where we are with current treatments.
On the drug discovery end, I think the model has the potential to be very useful. First, it identifies cognitive and negative symptoms as the 2 principal unmet therapeutic needs and, thus, gets the field to turn its attention toward treatment discovery for those components. Second, it calls attention to the need for development of alternative animal models, alternative screening, and novel pathways for therapeutic effect. In other words, it calls for an entirely different pathway from the one used in the development of antipsychotic medication. The models used to faithfully give us a drug that acts at the D2 dopamine receptor, with high predictive validity for its being an antipsychotic, do not get us drugs that are effective for the other 2 domains of schizophrenia.
This interview is published in collaboration with NARSAD: The Mental Health Research Association, and is supported by an educational grant from Pfizer.
Editor's Note:
The single-disease paradigm does not fully explain the pathophysiology of schizophrenia, nor does it allow for the most comprehensive treatment of patients with this condition. On behalf of Medscape, Randall White, MD, speaks with William T. Carpenter Jr, MD , Professor of Psychiatry and Pharmacology, University of Maryland School of Medicine, Baltimore Maryland, about why the etiology and treatment of cognitive and negative symptoms must be considered as individual domains separate from psychosis.
Medscape: Is it correct to say that, in your view, the single-disease paradigm of schizophrenia is flawed and has brought us to a nosological impasse?
William T. Carpenter Jr, MD: Yes, I sort of think that, but I'd put it a little bit differently. I would say that the single-disease entity has not been confirmed, and there are many reasons to have doubts about this idea and to expect that there is more than 1 etiologically based disease within the syndrome. The single-disease paradigm presumes that the different aspects of pathology are coming from the same latent structure and, therefore, would be similar both in etiology and in treatment response. The idea of a single disease entity has been a very weak heuristic. It has led to a focus on psychosis and a relative neglect of cognition impairments and negative symptoms. We do know that for treatment considerations, those are remarkably different aspects of pathology.
Medscape: You advocate looking at schizophrenia as having several domains of psychopathology. What are those?
Dr. Carpenter: This idea came from the international pilot study conducted years ago when I was working with John Strauss and John Bartko. We found that the measures of psychosis did not predict either how much of the other aspects of illness a person would have (like negative symptoms and interpersonal pathology) or future [outcome]. Each of these domains had its own history, its own nature of manifestation, and its own future. You could predict within, but not across, a domain.
The concept has been refined since then and includes cognition impairments, the type of impairments that you pick up on neuropsychological test assessment. It could also be that disorganization of thought and behavior is different from the reality-distortion symptoms in an important way. Rather than studying schizophrenia as a single disease -- expecting that 1 aspect of the condition would generalize to the others -- each of these domains should be studied in its own right in terms of both cause and treatment.
Medscape: What further research is needed to test the paradigm of multiple domains? And, are these domains independent?
Dr. Carpenter: As a construct, it's best to act as though they are independent and leave, for some future resolution, [the question of] why they co-occur in the illness syndrome. I think we'll be much more likely to understand etiology, pathophysiology, and make treatment discoveries if we make each domain its own target.
With that said, I doubt if these domains are truly independent. It could be that, for example, 1 disease process causes both psychosis and negative symptoms, while another disease process causes psychosis without negative symptoms.
Medscape: You have encouraged a distinction between negative symptoms and deficit symptoms. Can you explain?
Dr. Carpenter: Negative symptoms, probably best characterized as restricted affect, alogia, diminished drive, anhedonia, and avolition, can occur for a number of reasons. When first discussed, negative symptoms were meant to be direct manifestations of the schizophrenia itself. But, as the rating scales that captured them came into use, it turned out that people were rating negative symptoms regardless of whether they were caused by schizophrenia or not. For example, if a person is depressed, he or she will be anhedonic, and when over the depression, will be less anhedonic or may not be anhedonic at all. The state-dependent negative symptom is very different from the concepts developed by Kraepelin and Bleuler, who described a reduced capacity for these ordinary behaviors as a trait phenomenon in schizophrenia.
As another example, someone who is paranoid and socially withdrawn is rated as having negative symptoms. As she becomes less paranoid and, therefore, more socially engaged, it looks like her negative symptoms have improved, but it's all secondary to the [change in] paranoia. Similarly, a drug like haloperidol, particularly at higher doses, will induce negative symptoms. We introduced the term "deficit pathology" to refer to negative symptoms that are direct manifestations of the schizophrenia, not negative symptom ratings that can occur for a variety of reasons in people with schizophrenia but are not caused by the schizophrenia itself.
Medscape: Are there studies that permit an understanding of the long-term stability and prognosis of the deficit syndrome in schizophrenia?
Dr. Carpenter: We need to learn a lot more, but there are some things that we do know already. An aspect of this is a bit tautological. By definition, primary negative symptoms have to be present as trait pathology to meet the criteria. And, if the symptoms are present as trait, they're likely to be long-lasting into the future. So, the presence of primary negative symptoms will predict, in the long term, the presence of primary negative symptoms. If in fact there are rapidly fluctuating primary negative symptoms, we don't have a good way to detect them.
Medscape: Researchers such as Nancy Andreasen and Timothy Crow have suggested neurological features associated with deficit symptoms, like ventricular enlargement. Do you believe these findings have prognostic or therapeutic significance?
Dr. Carpenter: Well, in Crow's proposition, the difference between type I and type II schizophrenia would essentially be the presence or absence of irreversible negative symptoms. In that regard, it would be similar to how we think about deficit and nondeficit schizophrenia. I think his hypothesis that this was based on cell loss and would not be responsive to antipsychotic drug treatment did not prove to be correct.
Volume loss, particularly in the hippocampus, and increased ventricular volume are fairly widespread in schizophrenia. I don't believe they are restricted to the deficit subtype. In fact, in our studies, when we've isolated the deficit subtype, we've tended not to see as much volume loss as in the nondeficit form of schizophrenia. We've actually seen a tendency toward enlargement in other structures in the deficit form. It seems, therefore, that the neuropathology of schizophrenia differs between those with vs those without the deficit features.
Medscape: Enlargement in which structures?
Dr. Carpenter: In 1 study, there was enlargement in lobules 8, 9, and 10 of the cerebellum. In both normal [subjects] and nondeficit schizophrenics, the larger that structure, the better the performance on cognitive tests, and the nondeficits were not different in size from the normal controls. The deficit subgroup had larger lobules 8, 9, and 10, and, in this group, the larger that structure, the worse the cognitive performance -- as if there was a neuropathologic process that simultaneously made it larger and interfered with function.
Robert Buchanan's magnetic resonance imaging [MRI] studies tended to show greater volume losses in the nondeficit cases than in the deficit cases, and greater enlargement of the head of the caudate in deficit cases than in nondeficit cases.
Medscape: What do we know about long-term stability and prognostic implications of the cognitive dysfunction of schizophrenia?
Dr. Carpenter: Most, and perhaps even all, patients with schizophrenia have impaired cognition. It may be that some subgroups are spared but, in general, it looks like schizophrenia is associated with a reduction in IQ in the neighborhood of 5 to 10 points. It starts relatively early in life and is probably in place by the time people begin to manifest psychosis.
These impairments also appear to mark people at high risk for schizophrenia. It is possible that they begin, in some way, from birth, but the impairments certainly appear well in advance of psychosis and tend to be a stable trait throughout the course of the illness. You might conceptualize schizophrenia as an illness that begins not in late adolescence or early adulthood -- only the psychosis domain becomes manifest during that time period, while cognitive pathology manifests much earlier.
We think of cognitive impairment as trait pathology and as having an uncertain relationship with negative symptoms. Our studies and some others suggest that the deficit subgroup will have more impairment in tests that tap into frontal and parietal functions. But, basically, we just don't know the relationship between cognition and negative symptoms, largely because most studies have used measures that don't separate primary and secondary negative symptoms. There is a very low relationship between the cognition impairments and psychosis.
Medscape: What are the therapeutic implications of your domains-of-pathology model?
Dr. Carpenter: If you attempt to identify primary negative symptoms and look at the treatment studies to date, there's very little documentation of drug treatments having efficacy for primary negative symptoms. The negative symptom improvements that are routinely reported include secondary negative symptoms and may actually be restricted to secondary negative symptoms. The studies that have attempted to identify primary negative symptoms have not been able to document therapeutic efficacy of current pharmacotherapy.
Incidentally, any advantage in secondary negative symptoms is a true advantage, but theoretically it makes a huge difference as to whether or not you try to refine your antipsychotic drugs to treat negative symptoms of schizophrenia or whether it's an unmet therapeutic need that requires an entirely different discovery approach. The same can be said for cognition impairments. A number of drugs will look better on cognition and negative symptoms than the comparator haloperidol, but that may relate principally to the negative effects of haloperidol on those aspects of illness rather than the true efficacy for either cognition or negative symptoms of the newer treatments. So, the construct has helped us critique where we are with current treatments.
On the drug discovery end, I think the model has the potential to be very useful. First, it identifies cognitive and negative symptoms as the 2 principal unmet therapeutic needs and, thus, gets the field to turn its attention toward treatment discovery for those components. Second, it calls attention to the need for development of alternative animal models, alternative screening, and novel pathways for therapeutic effect. In other words, it calls for an entirely different pathway from the one used in the development of antipsychotic medication. The models used to faithfully give us a drug that acts at the D2 dopamine receptor, with high predictive validity for its being an antipsychotic, do not get us drugs that are effective for the other 2 domains of schizophrenia.
This interview is published in collaboration with NARSAD: The Mental Health Research Association, and is supported by an educational grant from Pfizer.
SHARE
