Neurocognitive deficits in schizophrenia are likely to be less severe and less related to the disorder than previously thought

Abstract

Even according to Kraepelin's observations more than one hundred years ago1, the term dementia praecox was an exaggeration, if not a misnomer. Not all of his patients showed signs of dementia, and a subgroup even recovered. Kraepelin also acknowledged that memory impairment, the core symptom of dementia, at times reflected lack of interest rather than faulty “impressibility of memory”. While the term was soon replaced by that of schizophrenia(s), the claim that the disorder is marked by global neurocognitive impairment lingers on. In fact, all major meta-analyses and reviews converge on the conclusion that patients with schizophrenia display large and global neurocognitive impairment, which many experts view as a core vulnerability factor for the disorder. Yet, the magnitude of these deficits – usually one standard deviation below the mean2 – is far less than the extent of impairments seen in patients with primary dementia, which is inconsistent with the idea that schizophrenia is a (praecox) form of dementia. We do not dispute that a large subgroup of patients show deficits on neurocognitive tests. Yet, we would like to emphasize that the degree to which these deficits can be attributed to schizo­phrenia itself is likely overestimated, whereas the degree to which they are due to medical and psychological factors that are often associated with schizophrenia, but that do not form part of the syndrome itself, has not been fully appreciated. These two classes of bias at times overlap, but should be distinguished for heuristic purposes. Importantly, this distinction is not just an academic issue. It has significant implications for: a) understanding why many people with schizophrenia appear to be cognitively impaired; b) understanding why the extent of observable cognitive impairment in people with schizophrenia can fluctuate widely depending on the environmental and interpersonal context; and c) choosing interventions to address the impairment. With respect to overestimation, an emerging literature indicates that poor performance is partly due to confounds during neurocognitive assessment. Most obviously, disorganization and derailed thinking, frequent symptoms in the disorder – especially under stress – may prevent proper understanding of task instructions and/or lead to avoidance of full engagement, with subsequent failure in more difficult tests. Patients may also experience interference from symptoms such as hallucinations, rumination and delusional ideas during assessment, further distracting them from the task at hand3. Moreover, patients’ motivation for assessment is often lowered, while anxiety and stress are higher compared to controls, and both of these factors are known to compromise performance3. According to Beck and colleagues4, 5, poor effort can explain one-quarter to one-third of the variance in test results. So far, the role of stigma related to diagnosis remains elusive. However, defeatist beliefs, which are a common consequence of being labelled with the diagnosis of schizophrenia, are associated with poor performance4, 5. Another potential source of bias that may contribute to an overestimation of deficits is the lack of representativeness of control participants. In some studies where psychiatric patients performed significantly worse than non-clinical controls, despite non-significant baseline differences on age or years of education, the performance of controls was in fact above average, rather than patients performing below average according to norm scores6. Apart from the aforementioned state factors compromising neurocognitive assessment, there are some conditions that can cause real and more lasting impairment, but do not reflect pathogenetic factors of schizophrenia. Perhaps most importantly, patients more often than controls suffer from obesity and diabetes (which can be independent of and/or related to treatment with antipsychotic medication) as well as cardiovascular disease (e.g., hypertension), and these factors compromise neurocognitive performance in both schizophrenia patients and the general population. In addition, hospitalization and loneliness have also been associated with poor neurocognitive performance7. One may argue that neurocognitive deficits were observed also in the pre-neuroleptic era and have been found even in at-risk individuals. However, the effects of, for example, defeatist beliefs (which are high in at-risk subjects, too) and hospitalization (at least in manifest patients) may have contributed to this. Poor neurocognitive performance may also represent an epiphenomenon of perceptual problems8. Furthermore, patients are often prescribed anticholinergic medications (clozapine or drugs aimed to reduce extrapyramidal symptoms) that compromise attention and memory. Finally, motor side effects, which are common with first-generation antipsychotics but can also occur with second-generation medications, can reduce performance on timed tests. Experienced neuropsychologists will be well aware of the aforementioned biases and confounds. In written individual reports, these may be acknowledged and perhaps even adjusted for. In group comparisons, however, these influences are traditionally ignored, as they are hard (e.g., motivation) or even impossible to control for (e.g., medication when the control group is not medicated at all). We advise researchers to either control/adjust for these factors where possible, for example through mediation analyses3, or to acknowledge possible sources of exaggeration of deficits in the abstract and discussion of their manuscripts. We also advise researchers not only to report mean values, but also the percentage of patients performing one and two standard deviations below the norm, to more fully describe the level of impairment in the sample. Often, only a minority of patients drives group differences that are then extrapolated to the entire population. Some aforementioned biases resulting in overestimation of deficits can be mitigated through creating a kind and motivating atmosphere during assessment9. In addition, it would help to test patients when distracting symptoms (e.g., hallucinations) are at a minimum. Change in medication might be sought, especially lower doses and minimization of drugs with anticholinergic properties. Addressing lifestyle factors related to hypertension, obesity and diabetes (e.g., weight loss, physical exercise) may reduce neurocognitive-relevant somatic risk factors, and there is tentative evidence that such interventions indeed enhance neurocognition10. Future consideration of these factors may open new windows for therapy beyond cognitive remediation, the traditional way to enhance neurocognition.

Topics & Concepts