Cognitive deficits are thought to be a core feature in schizophrenia and have been found to be strongly associated with impairments in functioning (
1). Neurocognition refers to the ability to correctly perceive, attend to, and remember information, such as verbal fluency, memory, processing speed, reasoning, problem solving, and attention (
2,
3). Social cognition refers to mental operations involved to understand, perceive, and interpret the social world, such as detecting emotion in faces and inferring what people are thinking and feeling (
2). In the literature, there is a tendency to conflate cognition and neurocognition. It is estimated that more than 70% of patients with schizophrenia have cognitive impairment (
4). There is an emerging literature on the course, associated factors, impact, and treatment of cognitive deficits in older adults with schizophrenia. Most of the latter has focused on neurocognitive deficits, although there have been a few recent interventional strategies targeting social cognitive dysfunction.
There is a pressing need to more fully understand the issues affecting cognition in older adults with schizophrenia because people age 55 years and older will soon represent one fourth or more of individuals with schizophrenia in many developed countries (
5). In this review, we address several questions. First, are there cognitive differences between older adults with schizophrenia and their age-matched peers? Second, are there cognitive differences between younger and older adults with schizophrenia or among persons with early- and late-onset disorders? Third, what factors are associated with cognitive functioning in older adults with schizophrenia, and how do they interact over time? Fourth, what are the later-life trajectories of cognitive functioning in older adults with schizophrenia? Fifth, are older adults with schizophrenia more prone to develop neurocognitive disorders? If so, how can clinicians distinguish the cognitive impairment of schizophrenia from such disorders? Sixth, what treatment strategies can be used to treat the cognitive problems of older adults with schizophrenia?
To address these questions, we have focused on more recent articles and have strived to reduce unnecessary detail and extensive referencing by citing conclusions from meta-analyses and literature reviews. There is no consensus definition of older adults with schizophrenia, but the literature has typically considered it to be age 55 years and older or, in some instances, age 50 years and older.
Cognitive Differences Between Adults With and Without Schizophrenia
Most broad-age range studies suggest that schizophrenia is associated with mild premorbid cognitive deficits that are approximately one third to two thirds of a standard deviation below age-matched peers in a control group (
6). However, a recent study of people age 50 years and older found that schizophrenia is associated with cognitive deficits that are nearly two standard deviations below cognitive levels of participants in a control group (
7). Rajji and Mulsant’s (
8) cross-sectional literature review of late-life schizophrenia (defined as schizophrenia, schizoaffective, or schizophreniform disorder among individuals age 50 years or older) found that patients with late-life schizophrenia consistently exhibited impaired global cognitive functioning compared with their age-matched healthy peers. Moreover, people with late-life schizophrenia demonstrated specific impairments in executive function, visuospatial ability, and verbal fluency, with impairment less consistently observed in memory, attention, and working memory. Subsequently, Irani et al. (
9) conducted a cross-sectional meta-analysis focused on older adults with schizophrenia (defined as 50 years of age or older, with most participants approximately 65 years of age). The authors found the largest effect in domains associated with language, immediate memory, and executive function. With respect to age of onset, Vahia and colleagues (
10) found that persons with early- or late-onset disorders had significantly worse cognitive scores than persons in a healthy control group.
Cognitive Differences by Age and Age at Schizophrenia Onset
Whereas the
DSM-5 (
11) and
ICD-10 (
12) do not distinguish by age of onset, the International Late-Onset Schizophrenia Group (
13) proposed that schizophrenia be termed “late-onset schizophrenia” for disorders that have an onset between the ages of 40 and 60 years, and “very-late-onset schizophrenia-like psychosis” for onset after the age of 60 years. The literature has sometimes ignored these demarcations and classified all people developing the disorder at age ≥40 (or 45) as one group. The type with onset between the ages of 40 and 60 years has been thought to be more similar to the early-onset subtype, although subtle differences have been noted between these two types, such as a preponderance of women, a lower level of symptom severity, and lower average antipsychotic dose requirement (
14). The very-late-onset type is distinguished by its much higher frequency of diagnosis among women than among men; greater prevalence of persecutory and partition delusions; higher rates of visual, tactile, and olfactory hallucinations; lower genetic load; the absence of negative symptoms or formal thought disorder (
13); and possibly a higher standard mortality rate versus older people with early-onset disorder, chiefly because of higher rates of comorbid illness and accidents (
15).
Rajji et al. (
16) conducted the most comprehensive study examining differences across age groups among people with schizophrenia. They compared people with schizophrenia, age 19–79 years, with their healthy age peers. There was a significant age effect in the group with schizophrenia, with progressive declines in overall cognitive scores as well as all cognitive domains except social cognition. The rate of decline across the various age groups was similar to that observed in the healthy groups. Rajji et al. noted that “deficits occur around the time of onset and do not progress beyond what is observed due to effect of aging over the entire span of adult life” (p. 6). They concluded that although individuals with schizophrenia decline at the same rate as those without the disorder, they cross the threshold of clinical impairment earlier and, thus, exhibit premature aging.
The data are in conflict with respect to cognitive deficits in older patients with early-onset schizophrenia and late-onset schizophrenia. Studies have shown late-onset schizophrenia to have lower, greater, or equivalent levels of cognitive impairment versus early-onset schizophrenia. Some investigators have suggested that there may be a few differences in the cognitive deficit patterns between early- and late-onset schizophrenia that are independent of age or duration of illness, although any definitive conclusions are constrained by methodological differences across studies (
17,
18). A serious methodological confounder is the possibility that organic-related psychosis was included among samples with late-onset schizophrenia (
17,
18). Although acknowledging the methodological shortcomings and a paucity of comparative studies, Rajji and Mulsant’s (
8) literature review found that “all the cross-sectional and most of the longitudinal studies were unable to distinguish patients with early-onset schizophrenia from those with late onset in terms of their cognitive profiles” (p. 138). In a large comparative study (N=110 patients with late-onset schizophrenia; N=744 patients with early-onset schizophrenia), Vahia and colleagues (
10) found that when duration of illness was considered, the only differences between early- and late-onset schizophrenia were in processing speed and perceptual organization, which were more impaired in the former. Rajji and Mulsant (
8) speculated that the lack of differences in cognitive functioning between people with early- versus late-onset schizophrenia—despite having experienced dissimilar durations of illness—suggests differences in the sequence of cognition and psychotic symptoms in the two groups; that is, cognitive decline precedes psychosis in late onset, whereas it follows psychosis in early onset. Finally, two recent studies have found better social cognition scores linked to late-onset schizophrenia and that appeared associated with age of onset rather than age per se (
19,
20).
Later-Life Trajectories of Cognitive Functioning in Schizophrenia
Investigators have been divided into two schools of thought regarding the lifetime course of cognition in schizophrenia: neurodevelopmental versus neurodegenerative patterns. To date, there has been no resolution of this debate (
19). This debate is further complicated by emerging evidence for longitudinal heterogeneity (
25,
26); that is, different patterns are observed among individuals with the disorder. The neurodevelopmental process suggests a pattern and rate of cognitive change with aging that appears parallel to that seen in the general population (
6,
8,
9,
24,
27,
28,
30,
40,
44). Mild cognitive impairment starts around the onset of the disorder and persists throughout life and across ages (
6,
8,
9,
24,
27,
28,
30,
40,
44). Bonner-Jackson and colleagues (
29) found that the most dramatic change in cognition was observed at early follow-up assessments within the first 7.5 years. Presumably this cognitive change is close to the index psychotic event, whereas after this period differences in the rate of cognitive decline between people with schizophrenia and controls diminish. This suggests that, except for the typical declines with aging, cognition is somewhat stable in patients with schizophrenia over time after the acute psychotic phase (
29). Thus, the neurodevelopmental model is frequently linked to the “static encephalopathy” model because it postulates that there are no marked declines in cognition in the years following the onset of the disorder. A recent meta-analysis by Heilbronner and colleagues (
48) provided further confirmation for this pattern. Moreover, a longitudinal study by Schnack and colleagues (
49) comparing magnetic resonance imaging scans of healthy participants and people with schizophrenia (age range = 16–67 years) found that there was an accelerated pattern of brain aging in people with schizophrenia at illness onset and during the initial five years; however, after that period, the rates of decline were comparable between the two groups, although the brain gap stayed constant. Unfortunately, long-term longitudinal neuroimaging studies are not available.
Neurodegenerative cognitive changes were postulated in the late 19th century by Kraepelin’s characterization of the disorder as “dementia praecox” (
50). A meta-analysis by Shah and colleagues (
51) examining evidence for cognitive decline late in the disorder found a ratio of positive to negative studies of 1.2:1 when control groups were used. In general, investigations of older adults with schizophrenia have found that cognitive deficits may be affected by institutional status, whereas age of onset may not contribute to cognitive functioning. For example, Harvey (
50) identified two separate periods of deterioration in schizophrenia patients: the first occurring some time prior to the first psychotic episode and the second beginning at approximately 65 years of age. Harvey and colleagues (
45) identified long-term institutionalization as a predictor of late-life decline; however, even after deinstitutionalization, patients continued to worsen cognitively. It was estimated that these patients who were chronically institutionalized—who represent a minority of older adults with schizophrenia—declined cognitively after the age of 65 years by 1 point per year on the Mini-Mental State Exam (MMSE) compared with 3 points per year for patients with Alzheimer’s disease (
30). Rajji et al.’s (
17) review of older patients with schizophrenia concluded that the findings of cognitive decline at age 65 years found in people who are institutionalized might not apply to those who have been living in the community. Indirect support for this hypothesis came from a study by Nemoto and colleagues (
26) who found that patients who were chronically institutionalized (mean age = 55 years) showed a small degree of improvement in some cognitive deficits, social functioning, and symptom burden after being transitioned into the community. This finding confirmed the importance of the connection between residential setting and cognition, although this conclusion differed from Harvey’s group, who had found that cognition continued to decline in discharged patients (
45).
Increasingly, there has been interest in potential for within-group improvement or heterogeneity in cognitive functioning in older adults with schizophrenia. Thompson and colleagues’ (
25) study of middle-aged and older people with schizophrenia (age 40–100 years) with a mean of 3.5 years of follow-up demonstrated that three subgroups exist with differing trajectories of cognitive ability. One group exhibited relatively high and stable trajectories of cognition (50%); a second group exhibited lower, modestly declining trajectories (40%); and a third group exhibited lower, more rapidly declining trajectories (10%). Even in the latter category, rates of cognitive decline were substantially less than people with Alzheimer’s disease. Thompson et al. observed that their findings suggested “a middle-ground between studies of community dwelling outpatients indicating stability and studies using data from institutionalized patients indicating accelerated decline” (p. 95). A somewhat similar heterogeneous pattern was observed in a more recent 52-month follow-up study of 104 community-dwelling people age 55 years and older living in New York City (
52). The authors found that 19% of the participants were rapid decliners, 25% were modest decliners, and 56% were stable or improved. Indeed, a notable finding was that 20% of the sample showed a .50 or greater effect size improvement in their cognitive scores (
52). There was flux in scores across all cognitive subscales, but memory and executive function showed the most fluctuation. A study by Savla and colleagues (
24) similarly found that over a 15-month period, as many participants cognitively improved as declined (about one sixth in each category) among their sample of middle-aged and older people with schizophrenia (mean age = 53 years). Other studies have alluded to heterogeneity in cognitive outcome within older adult populations with schizophrenia (
38). This issue was thoughtfully addressed by Shmukler and colleagues (
4) in their literature review:
We accept that the heterogeneous groups of patients may show divergent dynamics of cognitive functioning over time, in particular at different stages of schizophrenia: cognitive deficit in a subset of patients may remain stable throughout the illness, in other cases, cognitive functioning may improve and yet other patients may exhibit the progressive worsening of cognitive deficit. It is also necessary to highlight the possibility of undulating dynamics of cognitive impairment, with periods of deterioration and improvement (for example, due to relapse or remission of symptoms). In addition, it is necessary to take into account the possibility of divergent dynamics of the distinct cognitive domains in the same patient, thus requiring that when single patients are being assessed, the assessor should specifically look at separate cognitive domains. (p. 1008)
Do Patients With Schizophrenia Develop Neurocognitive Disorders?
The
DSM-5 (
11) made the decision that the cognitive impairment accompanying schizophrenia is different from Neurocognitive Disorders, which have more clear neurological etiologies. Section III of the
DSM-5 contains criteria for assessing the level of cognitive impairment. However, this decision on classification has created diagnostic conundrums, especially when older adults with schizophrenia experience declines in cognitive functioning.
Because of the increased propensity for patients with schizophrenia exposed to second-generation antipsychotic medications to develop metabolic syndrome, which is a risk factor for both Alzheimer’s disease and vascular dementia, it is reasonable to postulate that older adults with schizophrenia might be more likely to develop neurocognitive disorders. Nonetheless, the data in support of this notion have not been compelling. First, the only available study that has investigated the prevalence of metabolic syndrome in older patients with schizophrenia did not report a higher prevalence than their age peers (
53). Second, estimates of dementia are confounded by the fact that people with schizophrenia have cognitive deficits arising earlier in the disease course that are progressively worsened by typical age-associated cognitive decline. Consequently, as many as half of the people with schizophrenia may meet the cognitive cutoffs for mild dementia (
54). For example, a large-scale study using outpatient data by Hendrie et al. (
55) (mean age = 70 years) reported that patients with schizophrenia were diagnosed as having dementia twice as much as patients without schizophrenia (64% vs. 32%). However, Hendrie et al. conceded that
The lack of a significant effect on mortality of the diagnosis of dementia in the patients with schizophrenia . . . in contrast to the increased effect of a dementia diagnosis in the total sample . . . raises the possibility that at least in some cases physicians are misidentifying patients in these older, difficult-to-evaluate, patients with schizophrenia. (pp. 7–8)
There are few longitudinal studies examining the incidence of dementia in older people with schizophrenia. Shah and colleagues (
51) found only three longitudinal studies comparing people with schizophrenia with a nonschizophrenia control group; all were conducted on people with late-onset disorder, and follow-up ranged from three to 10 years. Two of the studies reported an increased prevalence of dementia, and one did not. None of the studies controlled for risk factors or noted the types of dementia that arose.
Friedman and colleagues (
30) observed that the cognitive decline experienced by people with schizophrenia who were chronically institutionalized and who showed accelerated declines in cognition beginning at age 65 years is distinctive in its course compared with Alzheimer’s disease. Studies of patients with Alzheimer’s disease have demonstrated a mean annual rate of decline on the MMSE between 2 and 5 points annually, whereas the patients with schizophrenia had a mean decline of 1 point annually over a six-year follow-up study. A longitudinal study by Palmer and colleagues (
44) found that there were comparable cognitive changes for patients with late-onset schizophrenia, patients with early-onset schizophrenia, and a control group of persons without schizophrenia, whereas the groups with Alzheimer’s disease had significantly greater declines. Moreover, in postmortem examinations of older adults with schizophrenia, it has been noted that there is minimal neurodegenerative brain pathology typical of Alzheimer’s disease pathology such as neurofibrillary tangles or senile plaque formation (
56).
“Very-late-onset schizophrenia,” which develops after age 60 years, is a more disputed domain with respect to outcome. On the basis of one study, a higher proportion of patients with very-late-onset schizophrenia versus their healthy age peers may go on to develop a type of dementia by 3- to 4.6-year follow-up, although incidence found was small (4.40% vs. 2.15%) (
57). However, it remains unresolved as to whether late-onset psychosis is a triggering factor or a prodromal form of dementia; that is, it’s essentially a misdiagnosis (
58). Biological markers of schizophrenia with a later onset have been elusive, and no specific neuropathological substrate has been identified (
10). Hahn et al.’s (
59) review of neuroimaging studies that compared early-onset and late-onset schizophrenia reported differences in eight of 10 studies, but these were not consistent across studies.
Increased microglial activity and neuroinflammatory processes have been implicated in both schizophrenia and various neurodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease (
60). It is unclear whether this might place older adults with schizophrenia at greater risk for such neurodegenerative disorders. Interestingly, a small study by Takano and colleagues (
61) found an association between microglial activation and both age and duration of illness among persons with schizophrenia.
Polygenic Risk Scores may in part explain the differences in disease processes. “A Polygenic Risk Score is the sum of trait-associated alleles across several genetic loci, typically weighted by effect sizes derived from a genome-wide association study” (
62). Liebers et al. (
63), who conducted an 8616 genotyping and clinical study of middle-aged and older adults followed over an average of 10 years, found “evidence for different degrees of association between polygenic risk for SCZ and genetic risk factors for ALZ on cognitive function and decline, highlighting potential differences in the pathophysiology of cognitive deficits seen in SCZ and ALZ” (p. 984). The presence of the established
APOE4-
TOMM40 risk loci was strongly associated with lower cognitive scores of patients with Alzheimer’s disease, particularly the subsets of language, attention, and verbal memory. Meanwhile, schizophrenia polygenic risk alleles are associated with general modest cognitive decline specifically in attention and spatial working memory. In summary, a preponderance of evidence—histopathological, clinical studies, and polygenic risk scores—suggests real differences in the genetic susceptibility, cognitive limitations, and disease progression of older adults with schizophrenia in comparison with people with Alzheimer’s disease.