Background
Pharmaceutical agents acting primarily on Dopamine-2 (D2) receptors (‘antipsychotics’) are a cornerstone of the current medical treatment for psychotic disorders. However, their side effects often pose a significant burden on patients (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; Read & Sacia, Reference Read and Sacia2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020). Ecologically valid insights into different side effects across various social contexts are important for well-informed decision-making between clinicians and patients. However, the currently available evidence from side effects as they are experienced in daily life is scarce (Bos, Schoevers, & aan het Rot, Reference Bos, Schoevers and aan het Rot2015).
There is robust evidence (Ceraso et al., Reference Ceraso, Lin, Schneider-Thoma, Siafis, Tardy, Komossa, Heres, Kissling, Davis and Leucht2020; Haddad, Das, Keyhani, & Chaudhry, Reference Haddad, Das, Keyhani and Chaudhry2012; Zhang et al., Reference Zhang, Gallego, Robinson, Malhotra, Kane and Correll2013) showing that antipsychotic medication is effective for managing acute symptoms of psychosis. Most patients state that pharmaceutical intervention during the acute phase improved their recovery process (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020). However, once acute psychotic symptoms abate, side effects of antipsychotic medication can become more cumbersome (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; Read & Sacia, Reference Read and Sacia2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020) and fuel the frequently expressed wish (Kahn et al., Reference Kahn, Fleischhacker, Boter, Davidson, Vergouwe, Keet, Gheorghe, Rybakowski, Galderisi, Libiger, Hummer, Dollfus, López-Ibor, Hranov, Gaebel, Peuskens, Lindefors, Riecher-Rössler and Grobbee2008) to discontinue or taper the medication earlier than currently recommended in various national guidelines (NICE, 2014; Orygen, 2010; Veling et al., Reference Veling, Boonstra, van Doorne, de Haan, van der Gaag, Gijsman, Kamstra, Klaassen, Kleijwegt and Lansen2017). Qualitative research shows that, in addition to somatic side effects including extrapyramidal motor symptoms and metabolic disturbances, patients taking antipsychotic medication can experience psychological side effects (PSEs) such as blunted emotional experience, reduced motivation, difficulty concentrating, or tiredness (i.e. mental fatigue) (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; de Haan et al., Reference de Haan, Lavalaye, Linszen, Dingemans and Booij2000, Reference de Haan, van Bruggen, Lavalaye, Booij, Dingemans and Linszen2003; Read & Sacia, Reference Read and Sacia2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020; Wolters, Knegtering, Wiersma, & Van Den Bosch, Reference Wolters, Knegtering, Wiersma and Van Den Bosch2003). These PSEs have previously been labeled as ‘secondary negative symptoms’ (Kaar, Natesan, McCutcheon, & Howes, Reference Kaar, Natesan, McCutcheon and Howes2020; Kirschner, Aleman, & Kaiser, Reference Kirschner, Aleman and Kaiser2017; Mas et al., Reference Mas, Gassó, de Bobadilla, Arnaiz, Bernardo and Lafuente2013), since they resemble the ‘primary negative symptoms’ inherent to schizophrenia spectrum disorders (Correll & Schooler, Reference Correll and Schooler2020). They may generally be explained by suppression of dopamine activity or simultaneous action on other neurotransmitter systems (Kaar et al., Reference Kaar, Natesan, McCutcheon and Howes2020; Kirschner et al., Reference Kirschner, Aleman and Kaiser2017; Sabe, Zhao, Crippa, & Kaiser, Reference Sabe, Zhao, Crippa and Kaiser2021). Whereas PSEs appear to be common (e.g. nonprimary negative symptoms were present in 45% of 1,452 people with psychotic disorders (Bobes et al., Reference Bobes, Arango, Garcia-Garcia and Rejas2010)), they remain largely underrepresented in the current literature on side effects of antipsychotic medication (Longden & Read, Reference Longden and Read2016) and have rarely been investigated in the flow of daily life (Bos et al., Reference Bos, Schoevers and aan het Rot2015). The resulting lack of ecologically valid insights into people’s emotional experiences under antipsychotic treatment impedes the understanding of the full side effect breadth, while this is crucial for providing optimal clinical advice.
Ecological momentary assessment (EMA) (or the experience sampling method [ESM]) is a method to study daily life emotions, experiences, and social context (Myin-Germeys et al., Reference Myin-Germeys, Oorschot, Collip, Lataster, Delespaul and van Os2009). Compared with retrospective clinical questionnaires, EMA allows for in-the-moment yet dynamic investigations (Granholm et al., Reference Granholm, Holden, Mikhael, Link, Swendsen, Depp, Moore and Harvey2019; Mote & Fulford, Reference Mote and Fulford2020; Wright et al., Reference Wright, Browne, Skiest, Bhiku, Baker and Cather2021), while mitigating the risk of peak-end and recall bias (Horwitz, Zhao, & Sen, Reference Horwitz, Zhao and Sen2023; Urban, Charles, Levine, & Almeida, Reference Urban, Charles, Levine and Almeida2018). The real-time assessments of momentary mental states such as positive and negative affect (PA/NA) are well suited to capture experiences that may be indicative of the PSEs described above (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; Read & Sacia, Reference Read and Sacia2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020; Wolters et al., Reference Wolters, Knegtering, Wiersma and Van Den Bosch2003), for example, altered PA/NA intensity and variability as indicators of blunted emotional experience. They also allow for distinct investigations at various times of the day, which may have relevant implications regarding the potential impact PSEs might have on people’s daily life. For example, mental fatigue could be especially burdensome in the morning when people need to get ready for daytime activities (including work, college, appointments, or other obligations), while blunted emotional experience could interfere with activities that involve social interactions. To the best of our knowledge, so far only a few studies have employed EMA to examine associations between the use of antipsychotic medication and momentary affect (Lataster et al., Reference Lataster, Myin-Germeys, Wichers, Delespaul, van Os and Bak2011a, Reference Lataster, van Os, de Haan, Thewissen, Bak, Lataster, Lardinois, Delespaul and Myin-Germeys2011b; Oomen, Simons, Broekmans-Madikrama, & Marcelis, Reference Oomen, Simons, Broekmans-Madikrama and Marcelis2024). One study suggests that increased D2 receptor occupancy in people with psychotic disorders taking high-affinity D2 antagonists is associated with decreased PA and increased NA (Lataster et al., Reference Lataster, van Os, de Haan, Thewissen, Bak, Lataster, Lardinois, Delespaul and Myin-Germeys2011b). Another study found that switching to aripiprazole was associated with decreases in both PA and NA in six patients with schizophrenia (Lataster et al., Reference Lataster, Myin-Germeys, Wichers, Delespaul, van Os and Bak2011a). Insights into other PSEs are lacking. Moreover, those findings require replication and might be different for people who are in recent clinical remission from first-episode psychosis (FEP).
Our study aimed to analyze EMA data of people in clinical remission from FEP in order to investigate how D2 affinity and dosage of antipsychotic medication are associated with momentary PSEs throughout the day, that is, overall as well as separately in the beginning, during and at the end of the day.
Based on previous studies (de Beer et al., Reference de Beer, Wijnen, Wouda, Koops, Gangadin, Veling, van Beveren, de Haan, Begemann and Sommer2024; Lataster et al., Reference Lataster, Myin-Germeys, Wichers, Delespaul, van Os and Bak2011a, Reference Lataster, van Os, de Haan, Thewissen, Bak, Lataster, Lardinois, Delespaul and Myin-Germeys2011b; Veerman, Schulte, & de Haan, Reference Veerman, Schulte and de Haan2017), we hypothesized that higher D2 affinity and higher dosage of antipsychotics are associated with increased levels of momentary blunted emotional experience, avolition, and mental fatigue, and that effect sizes are larger in the beginning of the day.
Methods
Study population
Data for the current study were collected between September 2017 and May 2023 through baseline assessments of the ongoing Handling Antipsychotic Medication: Long-term Evaluation of Targeted Treatment (HAMLETT) study (Begemann et al., Reference Begemann, Thompson, Veling, Gangadin, Geraets, van ’t Hag, Müller-Kuperus, Oomen, Voppel, van der Gaag, Kikkert, Van Os, Smit, Knegtering, Wiersma, Stouten, Gijsman, Wunderink, Staring and Sommer2020). HAMLETT is a multicenter, pragmatic, randomized controlled trial in the Netherlands evaluating the effect of early (dis-)continuation of antipsychotic medication on various outcomes in people who are in remission from FEP. Participants are between 16 and 60 years of age, in clinical remission (Andreasen et al., Reference Andreasen, Carpenter, Kane, Lasser, Marder and Weinberger2005) from FEP for 3–6 months, have sufficient command of the Dutch language, and have provided written informed consent. Exclusion criteria include dangerous or harmful behavior and coercive treatment with antipsychotic medication during FEP. Further details can be found in the published study protocol (Begemann et al., Reference Begemann, Thompson, Veling, Gangadin, Geraets, van ’t Hag, Müller-Kuperus, Oomen, Voppel, van der Gaag, Kikkert, Van Os, Smit, Knegtering, Wiersma, Stouten, Gijsman, Wunderink, Staring and Sommer2020).
At baseline, participants were invited to take part in the EMA add-on study until a pre-determined sample size was reached. Note that there was a temporary recruitment-pause due to COVID-19 and that not every participant has been invited to take part in the EMA add-on study. All study procedures were in line with local and international ethical standards, including the Declaration of Helsinki. Ethical approval was issued by the Medical Ethical Committee of the University Medical Center Groningen in the Netherlands (HAMLETT protocol number: NL62202.042.17, trial registration EudraCT number: 2017-002406-12).
Measurements
Baseline assessments took place during a study visit in one of the three study centers, a home visit, or a study visit in one of the recruiting mental health care institutions, and were performed by trained researchers. The EMA included repeated self-report data collection over eight consecutive days.
Sociodemographic, premorbid, and general clinical information was collected through the Comprehensive Assessment of Symptoms and History (CASH) (Andreasen, Flaum, & Arndt, Reference Andreasen, Flaum and Arndt1992), a semistructured clinical interview developed for schizophrenia spectrum disorders. The CASH includes information on substance use and the severity of symptoms during FEP based on five 6-point Likert items assessing the global severity of hallucinations, delusions, bizarre behavior, apathy, and anhedonia. Baseline clinical symptom severity was assessed with the Positive and Negative Syndrome Scale (PANSS) (Kay, Fiszbein, & Opler, Reference Kay, Fiszbein and Opler1987). Subjective well-being in the past week was assessed using the Subjective Well-being under Neuroleptics Scale (Naber, Reference Naber1995), a 20-item self-report questionnaire with higher scores indicating better well-being.
Data on the type and dosage of antipsychotic medication were retrieved from the Dutch Foundation for Pharmaceutical Statistics Stichting Farmaceutische Kengetallen (SFK). The SFK collects prescription data from >98% of community pharmacies in the Netherlands. In case SFK-data were unavailable, medication information was reported by the participant or their treating psychiatrist. We grouped the medications based on D2 receptor affinity (Ki cut-off value = 20.0; Kaar et al., Reference Kaar, Natesan, McCutcheon and Howes2020) and mechanism of action (D2 antagonism vs. partial agonism) into the following categories:
-
1. Low-affinity antagonists: olanzapine and quetiapine
-
2. High-affinity antagonists: haloperidol, risperidone, and amisulpride
-
3. Partial agonists: aripiprazole and brexpiprazole
Dosages were converted into olanzapine equivalents according to the 95% effective dose method (Leucht et al., Reference Leucht, Crippa, Siafis, Patel, Orsini and Davis2020).
Momentary psychological side effects of antipsychotic medication
EMA took place up to ten times per day for eight consecutive days. The questionnaire consisted of 37 items and should be completed within a few minutes. Prompts were sent within time blocks of 1:20 h, with time intervals of at least 30 minutes. The software environment was provided by RoQua (https://www.RoQua.nl). Participants were briefed about the EMA procedure by trained researchers and contacted to provide further instructions when they did not complete at least one questionnaire within the first two days. Responses were rated on a Visual Analogue Scale (VAS) of 1–100. In line with the current literature suggesting a response rate cut-off of one-third to ensure adequate validity of EMA (Delespaul, Reference Delespaul1995), we excluded subjects with <27 completed questionnaires from the analyses. Questionnaires were also excluded when subjects took >20 minutes for completion.
EMA items (Table 1) were selected to assess momentary mental states indicative of PSEs. Lack of initiating social contact or failing to complete daily life activities were selected as proxies for avolition (Oorschot, Kwapil, Delespaul, & Myin-Germeys, Reference Oorschot, Kwapil, Delespaul and Myin-Germeys2009), while difficulty concentrating or tiredness (Wolters et al., Reference Wolters, Knegtering, Wiersma and Van Den Bosch2003) were selected as proxies for mental fatigue. Similarly, established EMA items assessing PA/NA (Bell et al., Reference Bell, Eisner, Allan, Cartner, Torous, Bucci and Thomas2023; Myin-Germeys & van Os, Reference Myin-Germeys and van Os2007; Reininghaus et al., Reference Reininghaus, Kempton, Valmaggia, Craig, Garety, Onyejiaka, Gayer-Anderson, So, Hubbard, Beards, Dazzan, Pariante, Mondelli, Fisher, Mills, Viechtbauer, McGuire, van Os, Murray and Morgan2016) intensity, stability, and variability were used to measure blunted emotional experience, in line with previous work (Bos, de Jonge, & Cox, Reference Bos, de Jonge and Cox2019; Hermans et al., Reference Hermans, Myin-Germeys, Gayer-Anderson, Kempton, Valmaggia, McGuire, Murray, Garety, Wykes, Morgan, Kasanova and Reininghaus2021; Lataster et al., Reference Lataster, Myin-Germeys, Wichers, Delespaul, van Os and Bak2011a, Reference Lataster, van Os, de Haan, Thewissen, Bak, Lataster, Lardinois, Delespaul and Myin-Germeys2011b). PSEs were investigated overall as well as separately for different times of the day, that is, beginning of the day (first three prompts per day), daytime (prompts 4–7) and end of the day (last three prompts per day). Cronbach’s alpha and McDonald’s omega coefficients were calculated to assess the internal reliability of the item groupings, when appropriate (Table 1).
EMA indicators of psychological side effects
Note: Higher scores indicate increased affect intensity and variability and more initiative of social contact and completion of daily life activities but less affect stability and mental fatigue. α, Cronbach’s alpha coefficient; ω, McDonald’s omega coefficient.
a The first prompt of the day was excluded from analyses.
Statistical analyses
First, we employed descriptive analyses to characterize the study sample and the distribution of PSE scores. This included the calculation of frequencies and percentages for categorical variables and means with standard deviations or medians with interquartile ranges for continuous variables. We also ran missing value analyses.
Associations between antipsychotic medication and momentary psychological side effects
We performed stepwise multilevel mixed effects linear regression analyses (maximum likelihood estimation [MLE]) to investigate serial cross-sectional associations between the D2 affinity and dosage of antipsychotic medication and momentary indicators of PSEs throughout the day. In a first step, we ran models including D2 affinity and dosage as the independent variables and PSEs as the dependent variable. When at least one of the predictors was statistically significant (at α = 0.05) in a given model, we added an interaction term between D2 affinity and dosage. When neither of the two predictors was statistically significant, we additionally ran two separate models including only D2 affinity or dosage, respectively, as the single independent variable. All models were adjusted for fixed effects of age, sex, tobacco, and cannabis use in the past month and symptom severity during FEP (CASH). Tobacco and cannabis use in the past month as well as symptom severity during FEP were included as potential confounders because they may be associated with PSEs of antipsychotic medication (Kumari & Postma, Reference Kumari and Postma2005; Rashmi et al., Reference Rashmi, Robin, Richard, Michael, Hitesh, Matthew, Robert, Philip and Sagnik2016). Intraclass correlation coefficients (ICCs) were calculated to assess the degree of interindividual variation in the associations. Analyses were conducted overall as well as separately for different times of the day, that is, in the beginning, during, and at the end of the day. For all analyses, we used Stata Statistical Software: Release 18 (StataCorp., 2021).
Results
In total, 85 out of 453 HAMLETT participants took part in the EMA add-on study. At baseline, 56 (66%) of those participants completed >26 EMA questionnaires and were currently taking antipsychotic medication, yielding a total of 3,005 questionnaires for our analyses. Thirteen participants (23%) were using psychotropic agents next to the antipsychotic, that is, benzodiazepines, anticholinergics, lithium, stimulants, or an SSRI/SNRI. The distribution of antipsychotic prescriptions was relatively equally spread (Table 2) and SFK data were available for 43 participants (77%). The average symptom severity (PANSS) at baseline reflects the HAMLETT inclusion criteria (i.e. being in clinical remission) as scores were generally low, especially in the positive subdomain (Mean [M] = 8.7, standard deviation [SD] = 2.5, Table 2). The global symptom severity during the FEP as, retrospectively, assessed with the CASH was also relatively low (M = 11.8, SD = 4.5, Table 2). Thirteen participants (23%) had at least one non-psychotic comorbid psychiatric diagnosis. The average EMA response rate was similar throughout the day and lower at night (66.2% between 6 and 13 h, 69.8% between 13 and 18 h, 66.8% between 18 and 24 h, 41.0% between 0 and 6 h), the average completion time per questionnaire was 2:32 min. Missing value analyses showed that there were no variables with >5% incomplete data (Table 3).
Sociodemographic and clinical sample characteristics (N = 56)
Note: FEP, first-episode psychosis.
a There were two missing values for this variable. For the regression analyses, these were complemented by data generated through multiple imputation using chained equations (MICE).
Within-person distribution of momentary psychological side effects
Note: Each score except for PA/NA variability was rated on a visual analogue scale of 1–100. Note that higher scores indicate increased affect intensity and variability and more initiative of social contact and completion of daily life activities but less affect stability and mental fatigue. NA, negative affect; PA, positive affect.
Regression analyses
Associations between D2 affinity and momentary psychological side effects
As compared with high-affinity D2 antagonists, the use of low-affinity D2 antagonists (B = 10.04 [95% CI: 0.59; 19.49], Table 4) and the use of partial D2 agonists (B = 12.98 [95% CI: 2.43; 23.53], Table 4) were associated with increased PA intensity. The difference in effects between low-affinity D2 antagonists versus partial D2 agonists was not statistically significant (B = 2.88 [95% CI: −6.94; 12.71]). There were no associations between D2 affinity and the other PSEs (Table 4). Except for the associations involving low-affinity D2 antagonists, results were largely consistent across different times of the day (Supplementary Materials 1–3).
Multilevel mixed-effects linear regression analysis of associations between D2 affinity/dosage and momentary psychological side effects (N = 56; 3,005 observations)
Note: Corrected for age, sex, symptom severity during first-episode psychosis (CASH), and current tobacco and cannabis use. Bold font indicates statistical significance at α = 0.05. Intraclass correlation coefficients (ICCs) originate from the final model, wherein either dosage alone or dosage and D2 group were included. Note that higher scores indicate increased affect intensity and variability and more initiative of social contact and completion of daily life activities, but less affect stability and mental fatigue.
Associations between dosage and momentary psychological side effects
A higher dosage of antipsychotic medication was associated with decreased intensity in PA (B = −1.11 [95% CI: −1.97; −0.24], Table 4, Figure 1), lower affect variability (B = −1.40 [95% CI: −2.52; −0.29], Table 4, Figure 1) as well as with greater instability in PA (B = 0.23 [95% CI: 0.04; 0.42], Table 4, Figure 1). Associations with PA stability were identified during the day and in the evenings but not in the mornings, while associations with affect variability were identified in the mornings and during the day but not in the evenings (Supplementary Materials 1–3). The ICCs in the models including PA intensity and affect variability were 0.55 (95% CI: 0.42; 0.68) (Table 4) and 0.65 (95% CI: 0.52; 0.76) (Table 4), respectively, while the ICC in the model including PA stability was 0.02 (95% CI: <0.01; 0.31) (Table 4). This indicates that intraindividual variation was relatively small in the former two and large in the latter model. There were no significant associations between dosage and NA intensity, NA stability, avolition, and mental fatigue (Table 4).
Positive affect (PA) intensity, positive and negative affect (PA/NA) variability and PA stability in relation to antipsychotic medication dosage (affect rated on visual analogue scales of 1–100).
Analyses on the interaction between D2 affinity and dosage indicated that associations between dosage and concurrent levels of PSEs did not differ across D2-affinity profiles.
Discussion
In the current EMA study, we investigated serial, cross-sectional associations between D2 affinity and dosage of antipsychotic medication and momentary mental states indicative of PSEs in 56 people in remission from FEP. Higher antipsychotic dosages were associated with reduced affect variability, lower PA stability, and decreased PA intensity. The latter was also associated with the use of high-affinity D2 antagonists.
Our analyses indicate that people who were taking haloperidol, risperidone, or amisulpride scored an average of 10–13 points lower on momentary PA (on a scale of 1–100) as compared with those taking olanzapine, quetiapine, aripiprazole, or brexpiprazole, and that a higher dosage could account for an additional reduction in PA, for example, 5.5 points for an increase equivalent to 5 mg olanzapine more per day. Furthermore, an increase of 5 mg olanzapine equivalents more per day was associated with a 7-point reduction in overall affect variability, indicating a measurable blunting of emotional experience associated with the use of antipsychotic medication. These results align with findings from an earlier EMA study in 109 people with chronic psychotic disorders showing that increased D2 occupancy was associated with lower PA in participants using high-affinity D2 antagonists (Lataster et al., Reference Lataster, van Os, de Haan, Thewissen, Bak, Lataster, Lardinois, Delespaul and Myin-Germeys2011b). In addition, our data suggest that higher antipsychotic dosages may be linked with more instability in PA, a phenomenon that has previously been identified for overall affect in FEP and at-risk populations as compared with a healthy control group (Hermans et al., Reference Hermans, Myin-Germeys, Gayer-Anderson, Kempton, Valmaggia, McGuire, Murray, Garety, Wykes, Morgan, Kasanova and Reininghaus2021). This could have implications with regard to altered emotional reactivity in psychotic disorders (Reininghaus et al., Reference Reininghaus, Kempton, Valmaggia, Craig, Garety, Onyejiaka, Gayer-Anderson, So, Hubbard, Beards, Dazzan, Pariante, Mondelli, Fisher, Mills, Viechtbauer, McGuire, van Os, Murray and Morgan2016), but the absence of associations with NA and the rather small effect size undermine the clinical relevance of those results. Taken together, the evidence may suggest that stronger D2 suppression by antipsychotic medication can lead to instable and generally lower levels of positive emotions, and that these might drive the frequently reported blunting of emotional experience (Bjornestad et al., Reference Bjornestad, Lavik, Davidson, Hjeltnes, Moltu and Veseth2020; Read & Sacia, Reference Read and Sacia2020; Thompson et al., Reference Thompson, Stansfeld, Cooper, Morant, Crellin and Moncrieff2020; Wolters et al., Reference Wolters, Knegtering, Wiersma and Van Den Bosch2003). Although the cross-sectional nature of our data precludes drawing causal conclusions, this interpretation is supported by an experimental randomized controlled trial wherein 85 healthy participants had increased levels of anhedonia under amisulpride compared with placebo (Berg, Riehle, Rief, & Lincoln, Reference Berg, Riehle, Rief and Lincoln2023). The results also align with the finding that a lower starting dosage of amisulpride was associated with a more favorable subjective response (de Haan et al., Reference de Haan, van Tricht, van Dijk, Arango, Díaz-Caneja, Bobes, García-Álvarez and Leucht2023) and corroborate the literature on the subjective tolerability of antipsychotic medication (Awad, Reference Awad2019) indicating, for example, that >70% D2 receptor occupancy is associated with poorer subjective well-being (de Haan et al., Reference de Haan, van Bruggen, Lavalaye, Booij, Dingemans and Linszen2003; Mizrahi et al., Reference Mizrahi, Rusjan, Agid, Graff, Mamo, Zipursky and Kapur2007).
Another HAMLETT study (de Beer et al., Reference de Beer, Wijnen, Wouda, Koops, Gangadin, Veling, van Beveren, de Haan, Begemann and Sommer2024) investigated associations between D2 affinity and clinically assessed negative symptoms at baseline after applying inverse probability of treatment weighting (i.e. thereby specifically addressing secondary negative symptoms). Participants using partial D2 agonists had significantly lower levels of secondary negative symptoms than those using D2 antagonists, and higher dosages were associated with more severe secondary negative symptoms across all D2-affinity profiles (de Beer et al., Reference de Beer, Wijnen, Wouda, Koops, Gangadin, Veling, van Beveren, de Haan, Begemann and Sommer2024). The similar observations from the current study support these findings by extending them to momentary affect as experienced in daily life.
It is worthwhile to note that the ICCs in the models investigating PA intensity and PA/NA variability indicate relatively large interindividual variation in their associations with antipsychotic dosages. This aligns with qualitative (Dijkstra et al., Reference Dijkstra, Rijkeboer, Noordhof, Boyette, Berendsen, de Koning, Bennen, Hofman and de Haan2024) as well as quantitative (de Haan et al., Reference de Haan, van Tricht, van Dijk, Arango, Díaz-Caneja, Bobes, García-Álvarez and Leucht2023) studies showing that the subjective response to antipsychotic medication can vary considerably between patients. In contrast to the variation across individuals, results were relatively consistent across daytimes, albeit with the greatest effect sizes throughout the day and in the evenings and smaller effects in the beginning of the day (Supplementary Materials 1–3). This could reflect the usual time of taking the medication, which is in the evening for antagonists, and side effects usually being most severe within 1–4 hours after ingestion. Alternatively, it might reflect that PSEs of antipsychotic medication can be superimposed by other factors influencing affect, for example, diurnal rhythm-related upward trends in PA in the mornings and downward trends in the evenings (Murray et al., Reference Murray, Nicholas, Kleiman, Dwyer, Carrington, Allen and Trinder2009) or possibly increased awareness of, or less distraction from, one’s emotional experiences in the evenings. Unfortunately, we do not have information on treatment compliance and timing of medication intake, which might have relevant implications for the interpretation of our findings, for example, in relation to the timing of peak plasma levels.
Considering our findings with respect to PA, it is somewhat unexpected that we found no association between the antipsychotic prescription and NA. This might suggest that dosage and D2 affinity are, in fact, distinctively associated with reduced experience of joy, or (consummatory) anhedonia (Visser et al., Reference Visser, Chapman, Ruiz, Raugh and Strauss2020). Indeed, levels of NA intensity were relatively low in our sample, indicating that depressive or anxiety-related mental states were, at most, of subclinical severity. This may have important implications for patients reporting reductions in their experience of pleasure, as anhedonia was found to impact personal recovery after FEP beyond symptomatic and functional recovery in another HAMLETT study (Maas et al., Reference Maas, Bohlken, Gangadin, Rosema, Veling, Boonstra, de Haan, Begemann and Koops2024). Overall, the current study contributes to mounting evidence highlighting a frequently overlooked psychological dimension of potential risks associated with antipsychotic-induced D2 suppression after FEP.
On a neurobiological level, our results could be explained by mesolimbic dopaminergic suppression (Juckel, Reference Juckel2016) with the potential risk of negative symptom exacerbation at commonly prescribed dosages (Sabe et al., Reference Sabe, Zhao, Crippa and Kaiser2021). Importantly, however, higher dosages were not significantly associated with any other PSEs we investigated. In fact, higher dosage and stronger D2 affinity were only associated with affective side effects but not with avolition or mental fatigue. One explanation might be that participants could have been less likely to complete the EMA questionnaire at moments when they experienced low motivation, difficulty concentrating or tiredness. Another explanation might lie in the concurrent use of other psychotropic agents or psychiatric comorbidity, but given the considerable variation in comorbidity and comedication among the relatively small number of affected participants, such an impact is likely to be minor. It is also possible that effects on those side effects are obscured by differences in simultaneous action on receptors such as histamine-1 (Fang et al., Reference Fang, Sun, Wang, Ren, Calabrese and Gao2016), which are, however, highly variable among the currently used antipsychotic medications. For instance, low D2-affinity agents have antihistaminergic properties as well (albeit to varying degrees), while high D2-affinity agents generally do not (except for risperidone). Nevertheless, a study using clinical assessment methods in 520 people with schizophrenia (Fervaha et al., Reference Fervaha, Takeuchi, Lee, Foussias, Fletcher, Agid and Remington2015) showed that sedative effects of antipsychotic medication were not associated with levels of motivation. Similarly, the authors found no association between motivation and the dosage or type of antipsychotic medication. It should be noted that dosages of antipsychotic medication in the HAMLETT baseline sample were generally low. Taken together, the current evidence indicates that levels of motivation in people who are in recent clinical remission from FEP are not associated with differences in the dosage or type of antipsychotic medications primarily targeting D2 receptors, at least when these agents are used in low dosages.
Strengths and limitations
This is the first study to implement digital EMA to investigate associations between antipsychotic medication and PSEs in daily life after clinical remission from FEP. Therefore, our findings contribute to and extend the current body of evidence concerning the impact of antipsychotic medication on subjective well-being. Findings may be helpful in clinical practice with respect to decisions about dosing, type, and continuation of antipsychotic medication after FEP (Karow et al., Reference Karow, Czekalla, Dittmann, Schacht, Wagner, Lambert, Schimmelmann and Naber2007). A particular strength is that our findings are based on a large number of data points originating from various daily life situations, including work as well as leisure contexts (Djordjevic et al., Reference Djordjevic, Jongsma, Simons, Oomen, de Haan, Boonstra, Kikkert, Koops, Geraets, Begemann, Marcelis and Veling2025), and were relatively equally distributed throughout the day, indicating valid representation of people’s daily life.
There are also limitations to our study. First, despite correction for severity of FEP based on the CASH, we cannot entirely exclude the possibility of bias related to confounding by indication (e.g. higher dosages for more severe illness). Nevertheless, symptom severity at the time of daily life assessments can be expected to have had, at most, a minimal impact on prescription patterns, as none of the participants had clinically relevant psychotic symptoms at the time of study inclusion. Second, due to the naturalistic study design, the medication groups were not equal in size and some associations between D2 affinity and PSEs might have therefore been missed. Similarly, future investigations will need to include larger and more diverse samples, for example, with respect to initial treatment response. Third, since our sample comprises only around one-fifth of all HAMLETT participants, the results may not be entirely representative of the total sample. Importantly, however, a recent study including the non-EMA HAMLETT sample identified similar patterns in associations between antipsychotic prescriptions and secondary negative symptoms as assessed through clinical questionnaires (de Beer et al., Reference de Beer, Wijnen, Wouda, Koops, Gangadin, Veling, van Beveren, de Haan, Begemann and Sommer2024), indicating that our results are most probably not specific to the EMA sample. Finally, although our questionnaire length and assessment frequency align with established EMA research protocols in psychotic disorders (Bell et al., Reference Bell, Eisner, Allan, Cartner, Torous, Bucci and Thomas2023), there may be bias related to fatigue or habituation effects (Ulitzsch et al., Reference Ulitzsch, Viechtbauer, Lüdtke, Myin-Germeys, Nagy, Nestler and Eisele2025). However, given that completion rates and response times were consistent with previous work (Bell et al., Reference Bell, Eisner, Allan, Cartner, Torous, Bucci and Thomas2023), the risk of such bias is small. Moreover, it is improbable that the respective bias would differ between D2 affinities.
Conclusion
In sum, the current study suggests that, after clinical remission from FEP, a higher antipsychotic dosage and the use of high-affinity D2 antagonists, as compared with partial D2 agonists or low-affinity D2 antagonists, are associated with flattened momentary affects and decreased PA in particular, but not with avolition or mental fatigue as estimated with EMA. It also suggests that blunted emotional experience associated with higher antipsychotic dosages might be mainly driven by those reductions in PA. The identified associations were largely consistent across different times of the day, though effect sizes related to PA were smallest in the mornings and greatest in the evenings. Future research using longitudinal EMA data of PSEs is warranted, since replication and extension of our results harbor implications for the medical management of psychotic disorders.
Acknowledgments
The authors would like to thank all participants, study site-specific dedicated includers, and our numerous student interns for their indispensable contribution to HAMLETT-OPHELIA. Collaborators of the HAMLETT-OPHELIA Consortium are Prof. Dr. Jim van Os, Prof. Dr. Filip Smit, Dr. Sanne Koops, Prof. Dr. Nico van Beveren, Prof. Dr. Nynke Boonstra, Bram-Sieben Rosema, Dr. P. Roberto Bakker, Dr. Sinan Gülöksüz, Dr. Joran Lokkerbol, Dr. Ben Wijnen, Dr. Bodyl Brand, Erna van ‘t Hag, Dr. Alban Voppel, Franciska de Beer, Sterre Kamphuis, Iris Hamers, Toon Scheurink, Jort Noorman, Prof. Dr. Therese van Amelsvoort, Dr. Maarten Bak, Dr. Steven Berendsen, Truus van den Brink, Dr. Gunnar Faber, Prof. Dr. Koen Grootens, Martin de Jonge, Dr. Henderikus Knegtering, Dr. Jörg Kurkamp, Prof. Dr. Gerdina Hendrika Maria Pijnenborg, Dr. Anton Staring, Dr. Natalie Veen, Dr. Selene Veerman, Sybren Wiersma, Dr. Albert Batalla, Ruben Curfs, Jan-Jaap Hage, Ellen Graveland, Joelle Hoornaar, Inge Hobus, and Dr. Karin Huizer.
Funding statement
The HAMLETT-OPHELIA project is financed by grants from the Dutch Medical Science Foundation ZonMW (#848041003 and #636340001) awarded to Principal Investigator Professor I. E. Sommer. ZonMW had no role in the study design, collection, management, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication. MD is funded by an internal research grant of the University Medical Center Groningen (MD-PhD grant number 21-40).
Competing interests
The authors declare none.
Ethical standard
Ethical approval was issued by the Medical Ethical Committee of University Medical Center Groningen (HAMLETT protocol number: NL 62202.042.17, trial registration EudraCT number: 2017-002406-12).
Open access
