Cognitive impairment and psychosis in schizophrenia: independent or linked conditions?

Like much else regarding the pathophysiology of schizophrenia, the relationship between cognitive impairment and psychosis is far from elucidated. An undisputed observation is that they frequently co‐occur. Green et al1 provide a scholarly overview of the state of knowledge in this area. However, they do not elaborate on potential reasons underlying the co‐occurrence of cognitive impairment and psychosis.

Much of what we observe as clinical neuroscientists is correlational. Severity of scores on the Positive and Negative Syndrome Scale (PANSS) correlate with social and vocational impairment, which in turn correlate with lifetime cumulative antipsychotic drug administration, which correlates with cardiovascular comorbidities. In health and disease, the correlation between two variables is, often, the sum of the genetic and environmental contributions to these effects.

Thus, one possibility is that cognitive impairment and psychosis have a shared etiology – genetic and/or environmental. The “shared etiology” hypothesis can be tested using various methods. Environmental shared etiology is often studied in epidemiological cohort studies. Environmental factors such as obstetric complications, abuse and trauma during childhood, drug abuse, and immigration have all been reported in relation to cognitive impairment, psychosis or both2. Genetic shared etiology has been investigated using behavioral genetic methods in twins and siblings, and more recently using molecular genetic methods.

Twin studies can measure the genetic correlation between two traits. A genetic correlation of 1 between trait A and B would imply that all of the additive genetic influences on trait A also impact on trait B. In one such study, the genetic correlation between schizophrenia and IQ was relatively high (r=0.75)3. However, as the correlation is only 0.75, close to half of the genetic variance in schizophrenia is actually independent of intelligence, suggesting that both traits have genes with specific effects. When specific cognitive domains rather than IQ were examined, the genetic correlations varied considerably. They were only 0.34 for tests measuring verbal knowledge and 0.79 for tests measuring working memory. In a second study, using a larger sample, the genetic correlation between schizophrenia and IQ was 0.46, suggesting that the majority of the genetic variance in schizophrenia is actually independent of intelligence3.

With advances in molecular genetics, polygenic risk scores – which reflect the impact of many risk alleles of small effect – can now be used to quantify the role of directly measured risk genes for schizophrenia on specific cognitive domains. Several studies examined the relation between schizophrenia risk genes and IQ in the general population. What these studies examine is whether genetic variants associated with increased risk for schizophrenia would be associated with poorer cognitive performance. Overall, the reported correlations between the schizophrenia polygenic risk score and IQ and working memory were less than 0.1 for childhood, early and late adulthood4, 5. Similar results were observed for a range of cognitive domains6. Looking at the association from the other direction, i.e. whether genetic variants associated with poorer cognitive performance would be associated with increased risk for schizophrenia, the magnitude of associations was small as well (<1% of variance)6.

Using bivariate genome‐wide complex trait analysis to estimate the amount of shared genetic factors between schizophrenia risk and cognitive ability, the genetic correlation between schizophrenia and IQ was 0.2, and that between schizophrenia and working memory was 0.19. In this study, there was no evidence supporting genetic overlap between schizo‐phrenia and measures of verbal knowledge or social cognition6.

A second possibility is that schizophrenia and cognitive impairment co‐occur because both are associated separately with a third factor. For example, studies have shown that schizophrenia is associated with structural brain abnormalities and that IQ is correlated with whole and gray brain matter volumes. It is therefore possible that the genetic correlation between schizophrenia and IQ is explained by the correlation of both traits with brain volume3.

It is also possible that cognitive impairment is neither an antecedent nor an enduring consequence of schizophrenia, but it represents non‐specific brain vulnerability. Since cognitive impairment characterizes almost all mental disorders, as well as neurological disorders, traumatic brain injury, and drug and alcohol abuse7, and is already present before the onset of most disorders, it seems to be a non‐specific indicator of brain vulnerability or brain malfunctioning.

A fourth and less investigated possibility is that schizophrenia and poor cognitive abilities co‐occur coincidently. A coincidental hypothesis of psychosis and cognitive impairment would posit the following. Most community‐based individuals with borderline‐low cognitive abilities, but without any other mental or emotional disturbances, never see a mental health professional and are never labeled with a psychiatric diagnosis. Similarly, persons with encapsulated delusions or hallucinations, which do not lead to disruptive behavior, live in the community and only rarely come to the attention of mental health professionals. It is only the concomitant manifestation of the two, affecting social and vocational functioning, which leads to help seeking and to a diagnosis of schizophrenia. This would be similar to an individual who suffers from congestive heart failure or an individual suffering from degenerative joint diseases. Neither would have serious mobility limitations, although the former would probably avoid effortful walking and the latter would probably use a walking stick. While the two conditions are pathophysiologically unrelated, their concomitant presence would probably lead to severe mobility impairment and probably necessitate the use of a wheelchair.

The probability of two independent events to occur in sequence is the product of the multiply of the probability of each event occurring separately. Hallucinations and delusions are acknowledged by a sizeable minority of the general population (3.2‐7.2%)8. Approximately one‐third of those people report a frequent occurrence of those symptoms. 13.6% of the population have IQ scores 1 standard deviation or more below the mean, a level of cognitive impairment similar to that consistently reported in schizophrenia9. Therefore, even if psychotic symptoms and cognitive impairment were independent of each other, it could still be expected that between 0.44% to 0.98% of the general population will experience psychotic symptoms and present with a cognitive impairment, which approximates the lifetime prevalence of schizophrenia of 0.7%10.

Overall, these results imply that, while the hypothesis of a shared genetic etiology for psychosis and cognitive impairment is not ruled out, the shared genetic liability is likely to be only modest. It is also important to consider that, like with any correlation, genetic correlations do not reveal directionality. Therefore, the path of causation cannot be inferred: lower IQ may increase liability to schizophrenia, or schizophrenia may cause lower intelligence3.