The switch from conventional to atypical antipsychotic treatment should not be based exclusively on the presence of cognitive deficits. A pilot study in individuals with schizophrenia

Cognitive disturbances are a core feature of schizophrenia and have been extensively studied in recent years [1]. Cognitive impairment is present before the onset of the illness [2] and is also found in healthy relatives of patients, although to a lesser degree [3]. In addition, this feature is not exclusively secondary to psychiatric symptoms or medication [4]. Cognitive impairment is a better predictor of future functional outcomes compared with positive symptoms [5‐7].

The positive action of conventional antipsychotics drugs (APDs) on cognition is considered mild or moderate [8] and is limited to certain cognitive domains such as sustained attention [9, 10].

Regarding novel antipsychotics, this supposed cognitive enhancement would be mediated by their capability to raise the level of dopamine and acetylcholine in prefrontal regions [11]. However, their different affinity for brain receptors may result in different procognitive profiles of each class of antipsychotics. Many studies support a cognitive enhancement of the different atypical antipsychotics: quetiapine and olanzapine [12], quetiapine and risperidone [13], ziprasidone and olanzapine [14]; olanzapine, quetiapine, and risperidone [15], risperidone and quetiapine focusing in schizophrenia with predominantly negative symptoms [16]. Moreover, this favorable effect on cognition seems to persist after controlling for confounding variables such as clinical state, learning, and cooperation [17]. Nevertheless, some authors [18] have reported a worsening in a working memory task in first episode schizophrenic patients after treatment with atypical antipsychotics.

Studies that attempted to demonstrate the association of a greater cognitive enhancement with novel versus conventional APDs are susceptible to biases and face several difficulties, which include confounding effects of clinical symptoms, previous and adjunctive medications (i.e., anticholinergics and benzodiazepines), and practice effects, especially when intervals between assessments are short [19, 20].

Many studies have compared the procognitive properties of conventional and atypical APDs. Results that stemmed from these various reports were compiled by Woodward et al. [21] in a meta-analysis of 14 typical/atypical comparative studies; these authors concluded that atypical APDs improve overall cognitive function to a higher degree than typical APDs.

Despite all these data, the benefits of atypical APDs over conventional APDs for the full range of cognitive disturbances in schizophrenia remain controversial, especially for the most severe impairments, i.e., serial learning, executive functioning, vigilance, motor speed, and verbal fluency [22]. Several authors suggest that the difference in potency between these two APD types, if any, is small [19]. Recently, results of the neurocognitive component of the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE), which is a NIMH-sponsored study, did not indicate a better procognitive profile of four atypical APDs when compared with the conventional APD perphenazine [23].

Most of the studies comparing typical versus atypical antipsychotics on cognitive performance are randomized blind trials. Since many clinicians consider changing from conventional to novel antipsychotics in order to enhance cognition, we think that naturalistic studies may provide a closer look to real clinical practice.

In the present naturalistic, retrospective study, we used a comprehensive neuropsychological battery of tests to assess the cognitive outcome in two groups of schizophrenic patients: the first group was treated with one or more conventional APDs over two years, while the second group was treated with one atypical APD. The treatments were not discontinued at any time and there were no switches in the type of APD administered. Patients' performance on tasks of executive functions, verbal working memory, short-term memory, verbal memory, visual memory, speed of processing, verbal fluency, and motor speed were assessed twice, two years apart.

The first objective of this study was to examine changes in cognitive impairment associated with long-term APD treatment. The second objective was to assess the effect of the type of antipsychotic treatment on change of cognition over two years and describe the potential differences between the groups.

The results from this pilot, naturalistic study suggest significant, but heterogeneous cognitive improvements for the total sample of schizophrenic subjects after two years of continuous antipsychotic treatment. Improvement was observed in the domains of semantic verbal fluency, executive functions, visual memory, and auditory immediate memory. Size effects of the differences were moderate with a more clinical relevance in the two verbal fluency tasks and visual/auditory memory (Rey figure immediate recall and Digit Span Test forward), where size effects were all over 0.6. In the opposite, after two years of treatment no cognitive improvements were observed in visual/motor processing, motor speed and working memory tasks.

As patients did not participate in psychotherapeutic interventions or programs aimed at improving cognitive function and considering the absence of significant differences between T1 and T2 in the PANSS scores, we suggest that these cognitive gains may be in part the result of continuous exposure to antipsychotic medication. Moreover, the characteristics of the sample in the present study correspond to clinically stable, chronic patients, previously stabilized by the ongoing antipsychotic medication, so the effects psychopathological changes on cognition, although not measured, are probably limited.

The relatively long time interval between the baseline and the endpoint would rule out, at least in part, learning effects associated with repeated testing [39, 20].

The cognitive improvement observed in our cohort supports the results obtained in the neurocognitive component of the CATIE trial [15] and in a recent naturalistic study [40], which reported an improvement in cognition after six months of continuous treatment with APDs. The positive change described in the present study seems to exclude a potential deleterious effect of conventional APDs on a wide range of neuropsychological parameters even at the moderate-to-high dosages used here (CPZ units = 1116). As the procognitive advantages of atypical APDs reported in comparative studies, which could be caused by the use of high dosages of the drugs administered in the conventional APD group [8], remain controversial, the absence of differences in the present study can be considered more robust. Nevertheless, high dosages of conventional antipsychotics may probably impair cognition as dose reduction in patients with high dosages may lead to cognitive improvement, as shown recently [41].

Our results do not support the hypothesis of a better cognitive outcome in patients treated with atypical APD. The variable "type of APD" did not predict the improvement of performance in any of the neuropsychological tests, as assessed using regression analyses; however, intake of anticholinergic drugs seems to predict cognitive changes significantly. These results support other studies that concluded that concurrent use of anticholinergic drugs, especially their acute administration [42], is another possible explanation for the negative effects of these drugs on cognition reported in the literature [39, 43]. In addition, this provides a further reason to avoid these drugs and the antipsychotics that require this sort of adjunctive medication to prevent extrapyramidal symptoms.

In contrast to the similar effect on cognitive evolution observed for both types of APDs, a significant improvement in the PANSS positive subscale score was detected in the group treated with conventional APDs. Interestingly, another significant effect, which translated into an improvement in the PANSS General Subscale score, was observed in the group treated with atypical APDs, which suggests a positive effect of these new drugs on the nonpsychotic symptoms of schizophrenia.

The limitations of the present study should be taken into account. Firstly, the relatively small sample size and the consequent lack of statistical power may have masked possible differences between the groups. In addition, these factors did not allow us to perform head-to-head comparisons of specific APDs. Secondly, as this was not a blind trial and patients were not randomly assigned to medication groups, the differences detected among groups may simply reflect a prescription bias. According to a recent meta-analysis [39], the differences in cognitive outcomes observed between conventional and atypical APDs in naturalistic, open-label studies are usually not found in double-blind randomized trials. Nevertheless, naturalistic studies have the advantage of using doses of APDs that are closer to those applied in daily practice, when compared with randomized clinical trials. The segregation of different conventional and atypical APDs into two groups represents the third limitation of this study. Although this procedure may make sense based on the clearly different mechanisms of conventional versus atypical APDs, it is also true that some intragroup differences may exist, especially among subjects taking different atypical APDs [44, 12]. Finally, the cognitive improvements observed for the schizophrenic patients after two years of treatment could be questioned, as the present study lacks a control group of healthy subjects. These limitations are in part compensated by the two-year lapse between the evaluations, which is substantially longer than that of most published longitudinal studies comparing conventional and atypical APDs. This long period of continuous treatment may help to minimize learning effects, which renders eventual differences between the groups more robust. Several animal/preclinical studies have demonstrated that both conventional and atypical APDs increase neurogenesis and proliferation of nonneuronal cells in the adult brain, particularly in some areas of the hippocampus (see 45 for a review). This may result in cognitive improvement in terms of memory and learning. The translational effect of neurogenesis on cognitive gains/enhancements can only be correctly evaluated by longitudinal studies that incorporate more than a few weeks between assessments.

Data from the literature showing that cognitive measures can predict functional outcomes [6] emphasize the importance of cognitive enhancement. Antipsychotic continuous treatment represents one of the few ways to remediate these deficits, albeit modestly. Therefore, strategies aimed at improving treatment adherence may help prevent cognitive decline and allostatic load [46].

Bearing in mind that this pilot study lacks a representative sample of patients, these preliminary finding are consistent with the recent literature in which atypical antipsychotics have not demonstrated cognitive benefits over typical or conventional antipsychotics. Our results did not show a clear advantage of atypical over conventional APDs on cognitive performance, however, the general improved tolerability profile of second generation antipsychotics regarding neurological side effects may facilitate treatment adherence, which in turn may result in cognitive improvement. Anyway, in the absence of other reasons to change ongoing treatment, which include negative or affective symptoms or lack of compliance with the regimen, the switch from typical APD at low or moderate doses to atypical APDs is not justified if based solely on the expectation of a more favorable cognitive outcome. Nevertheless, other psychopharmacological and psychosocial strategies should be implemented to enhance cognitive outcome in schizophrenic patients.