Abstract
Background:
Extrapyramidal symptoms (EPS) in association with the long-acting, injectable, atypical antipsychotic aripiprazole once monthly (AOM) have been observed in clinical trials, but information on EPS requiring treatment with anticholinergic drugs in clinical practice is limited.
Objectives:
The objective of this European post-authorisation safety study (PASS) was to assess the risk of EPS-related events, as defined by dispensings of anticholinergic drugs, linked to the use of AOM in routine clinical practice.
Design:
This European cohort study was based on healthcare databases from Germany (German Pharmacoepidemiological Research Database GePaRD), Italy (Caserta and Palermo healthcare claims databases), and Sweden (National health registers). New users of AOM were followed from their first dispensing for a maximum of 2 years.
Methods:
Primary study outcome was an EPS-related event, defined as the first dispensing of an anticholinergic drug. The crude incidence rate (IR) and the cumulative incidence of EPS-related events were estimated. Cox proportional hazard regression modelling was performed to further investigate the effect of potential risk factors for the occurrence of EPS-related events.
Results:
A total of N = 1748 patients were eligible for inclusion into the primary study populations (Germany N = 629; Italy N = 519; Sweden N = 600). IRs of EPS-related events per 100 patient years were highest in Italy (IR = 18.4; 95% confidence interval (CI) 15.3–22.1), followed by Sweden (IR = 7.7; 95% CI 5.8–10.2) and Germany (IR = 3.4; 95% CI 2.4–4.6). Rates were highest during the first 30 days after treatment initiation. Cumulative incidences after 2 years of treatment were 27.8% (Italy), 11.5% (Sweden), and 10.0% (Germany). Diabetes and previous antipsychotic drug use were identified as risk factors for EPS-related events.
Conclusion:
In this observational study, incidence rates of EPS-related events, defined as the first dispensing of an anticholinergic drug during follow-up, were compatible with the known safety profile of AOM but showed substantial regional variation.
Trial registration:
EU PAS number EUPAS21056.
Keywords: anticholinergic drugs, aripiprazole, cohort study, depot antipsychotics, extrapyramidal symptoms, long-acting injectable antipsychotics
Introduction
Extrapyramidal symptoms (EPS) are one of the clinically most relevant adverse effects of antipsychotics. Based on the timing of EPS occurrence after initiation of antipsychotic drug treatment, EPS have been classified into acute and tardive symptoms. 1 Acute EPS include parkinsonism, acute akathisia and acute dystonia, 2 while tardive dyskinesia represents the most important form of chronic EPS. 3 It has been recognized that not only first-generation antipsychotics with high affinity to dopamine receptors, but also the majority of second-generation antipsychotics with atypical mechanisms of action can cause EPS. 3 The clinical management of EPS occurring in the context of antipsychotic drug treatment is challenging. While for acute EPS, such as parkinsonism or dystonia, treatment with anticholinergic drugs is an effective pharmacological approach, it may worsen chronic EPS (tardive dyskinesia). 1 In addition, anticholinergic drugs themselves have important potential side effects such as cognitive deterioration, urinary retention, dry mouth, and risk of glaucoma exacerbation, especially in elderly people. 3 According to a WHO recommendation, dose reduction and switching of antipsychotics should be preferred over treatment with anticholinergics. Only if these have proven ineffective, or when the EPS are acute or severe, short-term use of anticholinergics may be considered. 4
Aripiprazole once monthly (AOM) is a long-acting injectable formulation of aripiprazole, a second-generation atypical antipsychotic. An overview of systematic reviews reported that aripiprazole caused significantly fewer general extrapyramidal adverse events, less use of antiparkinsonian drugs and akathisia, compared with typical antipsychotic drugs or risperidone. 5 However, meta-analyses of clinical trials indicated that in comparison to the oral formulations of second-generation atypical antipsychotics, the respective depot formulations may carry a greater risk of extrapyramidal adverse events.6,7 In the pivotal randomized clinical trials of AOM, the proportion of patients with treatment-emergent adverse events related to EPS was 14.9% (ASPIRE study) 8 and 21.9% (ASPIRE EU study). 9
Rates of EPS in patients treated with AOM in routine clinical practice are unknown. Etminan et al. 10 reported on the risk of extrapyramidal adverse events in users of aripiprazole as compared to nonuse of antipsychotics. The study did not specify whether depot formulations were included, nor did it report absolute risks of EPS. When AOM was launched in the EU in 2014, a post-authorisation safety study (PASS) using automated healthcare databases was initiated to further assess the risk of EPS, which were considered as an important identified risk at the time of launch. As this study was part of the European Risk Management Plan of AOM, the methodological aspects of the study were discussed and agreed upon with the European Medicines Agency (EMA) before study initiation. The objective of this PASS was to further assess the risk of EPS-related events linked to the use of AOM in routine clinical practice. EPS-related events were defined via the first pharmacy dispensing of an anticholinergic drug in the primary study analysis.
Methods
The reporting of this study conformed to the Strengthening the Reporting of Observational Studies in Epidemiology statement. 11
Setting
This was a European, multinational cohort study, using longitudinal administrative healthcare databases from three European countries (Germany: German Pharmacoepidemiological Research Database (GePaRD), Italy: Caserta and Palermo healthcare claims databases, Sweden: National health registers). All databases are established data sources for pharmacoepidemiological research and have been used for several previous studies (details on databases are included in the Supplemental Material). New users of AOM (incident users as detailed below) between the country-specific market entry date of AOM and the end of the respective study inclusion period were identified. The start of AOM marketing in the respective country represented the beginning of the database-specific inclusion time period. Healthcare data is usually included in consecutive calendar time packages in the databases (e.g., every year). The number of AOM users in the databases was monitored, and if a sufficient number of AOM users was reached (i.e., at least 500 patients), the end date of the current database cut defined the end of the database-specific inclusion time period. To ensure a possible follow-up time of 2 years for all patients identified as AOM users during the inclusion time period, the analyses were only performed after 2 additional years of data were included in the database. The inclusion time periods were for Germany: between June 15, 2014 and December 31, 2015; for Italy: between July 01, 2014 and December 31, 2017; and for Sweden: between January 01, 2014 and December 31, 2015. The first observed dispensing date of AOM for a patient after market entry was defined as the index date. All included patients were followed up from the index date for a maximum of 2 years, until either the first EPS-related event (i.e., a first pharmacy dispensing of an anticholinergic drug); end of the maximum 2-year observation period; end of the AOM treatment episode (i.e., discontinuation of AOM); de-registration in the respective database; or death, whichever came first (see Figure S1 in the Supplemental Material).
Participants
Patients who fulfilled all inclusion criteria (incident dispensing of AOM during the inclusion time period; availability of a database-specific covariate assessment time period prior to inclusion, see Supplemental Material for details) and none of the exclusion criteria (diagnosis of Parkinson’s disease in the 5 years before index date; diagnosis of EPS in the 45 days preceding the index date) were included in the study and formed the full study population. From this full study population, patients who had a dispensation of an anticholinergic drug (Anatomical Therapeutic Chemical (ATC) groups N04AA, N04AB, N04AC) in the covariate assessment time before the index date and up to 1 day after the index date were excluded to build the restricted study population. This was necessary because the new use of anticholinergic drugs during follow-up defined the outcome of interest (EPS-related events) in the primary analysis. Thus, the restricted study population was the primary study population to increase the likelihood that (1) patients were not users of anticholinergic drugs already before the start of AOM treatment and (2) patients did not receive anticholinergic drugs prophylactically together with their first AOM dispensing.
Exposure and covariates
AOM was the main exposure of interest in this study. The person-time classified as exposed to AOM was identified via data on pharmacy dispensings of AOM. The duration of each dispensing was calculated as the number of dispensed vials (powder and solvent for one injection) multiplied by 30 days, assuming that dispensed vials are injected once per month according to the recommended injection interval. 12 A patient was considered as continuously exposed to AOM if no gap longer than 2 months (i.e., twice the standard injection treatment interval) occurred between the estimated end of the previously dispensed prescription and the dispensing date of the next prescription. In case of a treatment gap of longer than 2 months, the estimated end of the last dispensing before that gap was considered as the end of AOM exposure (see Figure S2 in the Supplemental Material).
To explore potential risk factors for the occurrence of EPS-related events in users of AOM, data on important variables13–18 were collected: Demographic characteristics; EPS-related events in medical history (including type of events); diabetes; depression; previous use of antipsychotics, as well as other concomitant psychotropic medications. Details on variable definitions and codes are included in the Supplemental Material.
Outcome
The primary study outcome was an EPS-related event, defined as the first dispensing of an anticholinergic drug during follow-up. In addition, pre-planned sensitivity analyses were performed that considered diagnoses instead of or in combination with a first anticholinergic drug dispensing for identifying EPS-related events: In a first sensitivity analysis, EPS-related events were identified via specific diagnostic codes (see Supplemental Material, Table S1) instead of a first anticholinergic drug dispensing. A second sensitivity analysis used the same approach, but this time with an extended list of diagnostic codes. In a third sensitivity analysis, EPS-related events required a first anticholinergic drug dispensing; a corresponding diagnostic code from the extended code list during the exposure to AOM; and one record of a dispensed anticholinergic drug occurring no more than 30 days apart from the diagnosis and between the index date and the end of the exposure to AOM. The first and second sensitivity analyses were performed in the full study population, as the prevalent use of anticholinergic drugs was not considered to be an issue in this analysis. The third sensitivity analysis was performed in the restricted study population.
Study size
It was assumed that the incidence of EPS-related events in the healthcare databases is close to the one observed in AOM pivotal clinical trials, where the pooled cumulative risk of EPS adverse events was 18.4% (based on the two pivotal studies of Kane 2012 (52 weeks placebo-controlled) and Fleischhacker 2014 (38 weeks active-controlled)).8,9 With a minimum sample size of 500 patients per database, the expected absolute precision of the estimated incidence for EPS-related events with a 95% confidence was estimated as 3.5% (half the confidence interval), which was considered appropriate for the objectives of this study. Quantitative pooling of the results was considered inappropriate due to methodological differences of the data sources (e.g., population-based national health registers in Sweden vs national statutory health insurance database in Germany vs regional healthcare claims databases in Italy) and expected regional variation of event rates 19 and was thus not considered in study size estimation.
Statistical analyses
Descriptive analyses of the study populations were performed and summarized. The crude incidence rate (IR) of EPS-related events per 100 patient years was estimated, using the number of patients with at least one EPS-related event during the follow-up in relation to the total patient time at risk. The cumulative incidence of EPS-related events was estimated by using the Aalen-Johansen method, 20 considering all-cause death as a competing event. Corresponding cumulative incidence function plots were produced. To further investigate the effect of potential risk factors for the occurrence of EPS-related events, IRs stratified by baseline characteristics were calculated and Cox proportional hazard regression modelling was performed.
Results
Patient characteristics
A total of N = 1748 AOM users were eligible for inclusion in the restricted study populations (Germany N = 629; Italy N = 519; Sweden N = 600), after applying the inclusion and exclusion criteria (Supplemental Material; Table S2). Patients from the three countries had similar demographic characteristics (Table 1). The proportions of patients with previous diagnostic records of EPS were 0% (Germany), 0.8% (Italy) and 1.2% (Sweden). The majority of patients had used antipsychotics during the 5 years before the index date, with the highest proportion observed in Germany (94.9%), and the lowest in Italy (79.4%). The proportion of patients with recorded use of oral aripiprazole before initiation of AOM was lower than 50% in all countries. The proportion of patients who were hospitalized during a 5-year period before the index date was lower in Italy (68.6%) than in Sweden (90.5%) and Germany (93.2%). A similar pattern was observed for the 1-year period before the index date.
Table 1.
Baseline characteristics of patients initiating treatment with AOM (restricted study population).
| Variable | Germany (N = 629) |
Italy (N = 519) |
Sweden (N = 600) |
|---|---|---|---|
| Age (years) | |||
| Mean (SD) | 40.1 (12.9) | 42.1 (12.4) | 39.5 (13.2) |
| Min–Max | 14–85 | 13–84 | 16–87 |
| Gender, n (%) | |||
| Men | 319 (50.7%) | 279 (53.8%) | 326 (54.3%) |
| Women | 309 (49.1%) | 240 (46.2%) | 274 (45.7%) |
| Unknown/missing | 1 (0.2%) | 0 | 0 |
| Previous diagnoses of EPS in the 5 years before the index date, n (%) | |||
| Yes | 0 | 4 (0.8%) | 7 (1.2%) |
| No | 629 (100.0%) | 515 (99.2%) | 593 (98.8%) |
| Diabetes in 5 years before the index date, n (%) | |||
| Yes | 48 (7.6%) | 56 (10.8%) | 53 (8.8%) |
| No | 581 (92.4%) | 463 (89.2%) | 547 (91.2%) |
| Antipsychotic use in the 5-year period before the index date | |||
| Yes | 597 (94.9%) | 412 (79.4%) | 544 (90.7%) |
| No | 32 (5.1%) | 107 (20.6%) | 56 (9.3%) |
| Oral aripiprazole use in 45 days before the index date | |||
| Yes | 155 (24.6%) | 144 (27.7%) | 252 (42.0%) |
| No | 474 (75.4%) | 375 (72.3%) | 348 (58.0%) |
| Antipsychotic use other than aripiprazole in 45 days before the index date | |||
| No | 262 (41.7%) | 418 (80.5%) | 318 (53.0%) |
| Only atypical | 263 (41.8%) | 57 (11.0%) | 179 (29.8%) |
| Only typical | 35 (5.6%) | 39 (7.5%) | 76 (12.7%) |
| Typical and atypical | 69 (11.0%) | 5 (1.0%) | 27 (4.5%) |
| Prior hospitalizations | |||
| Within 5 years before the index date | 586 (93.2%) | 356 (68.6%) | 543 (90.5%) |
| Within 1 year before the index date | 467 (74.2%) | 257 (49.5%) | 426 (71.0%) |
SD, standard deviation.
Incidence rates and cumulative incidences of EPS-related events
The incidence rates (IR) of EPS-related events, defined as a first pharmacy dispensing of an anticholinergic drug, per 100 patient years were highest in Italy (IR = 18.4; 95% confidence interval (CI) 15.3–22.1), followed by Sweden (IR = 7.7; 95% CI 5.8–10.2) and Germany (IR = 3.4; 95% CI 2.4–4.6) (Table 2). In all three countries, rates were highest directly after initiation of treatment with AOM and decreased thereafter. The lowest rates were observed for the second year of treatment.
Table 2.
Incidence rates of EPS-related events (restricted study population).
| Time period (days) | Incidence rates (95% CI) per 100 patient years | ||
|---|---|---|---|
| Germany (N = 629) |
Italy (N = 519) |
Sweden (N = 600) |
|
| 0–30 | 17.6 (8.0–33.3) | 56.7 (38.0–84.6) | 18.4 (9.6–35.3) |
| 31–91 | 6.0 (2.2–13.0) | 26.4 (15.9–41.2) | 10.7 (5.6–20.5) |
| 92–183 | 4.0 (1.5–8.7) | 19.6 (11.8–30.6) | 10.7 (5.9–19.3) |
| 184–365 | 3.7 (1.9–6.7) | 16.4 (10.8–23.8) | 8.3 (4.8–14.3) |
| 366–730 | 1.4 (0.6–2.8) | 10.6 (7.0–15.5) | 2.6 (1.2–5.8) |
| Total (0–730) | 3.4 (2.4–4.6) | 18.4 (15.3–22.1) | 7.7 (5.8–10.2) |
EPS-related events were defined as pharmacy dispensings of anticholinergic drugs.
CI, confidence interval; EPS, extrapyramidal symptoms.
For cumulative incidences as estimated by the Aalen-Johansen method, similar differences between countries were observed: the estimated cumulative incidences after 2 years of treatment were 27.8% in Italy, 11.5% in Sweden, and 10.0% in Germany. As the estimate was considerably higher in Italy than in Germany or Sweden, the cumulative incidence in Italy was additionally estimated for the following time points (post-hoc analysis): 6 months (13.1%); 12 months (20.2%); 18 months (24.7%). The corresponding cumulative incidence curves are included in the Supplemental Material (Figures S3, S4, and S5).
Association of baseline characteristics with the risk of EPS-related events, defined as pharmacy dispensings of anticholinergic drugs
The following variables were associated with increased risks of EPS-related events: Diabetes mellitus in the 5-year period before the index date (identified in the German database); Antipsychotic use in the 5-year period before the index date (identified in the databases of Italy and Sweden); and use of antipsychotic drugs other than aripiprazole in 45 days before the index date (identified in the Italian database). The crude incidence rates and the corresponding adjusted hazard ratios are reported in Table 3. The results for other variables (not associated with increased risks of EPS-related events) are included in the Supplemental Material (Table S3).
Table 3.
Incidence rates of EPS-related events (restricted study population), stratified by baseline characteristics, and adjusted hazard ratios for baseline characteristics.
| Baseline variable | Incidence rates (95% CI) per 100 patient years | Adjusted hazard ratio (95% CI) a | ||||
|---|---|---|---|---|---|---|
| Germany (N = 629) |
Italy (N = 519) |
Sweden (N = 600) |
Germany | Italy | Sweden | |
| Diabetes in the 5-year period before the index date, n (%) | ||||||
| No | 3.1 (2.2–4.3) | 18.6 (15.4–22.5) | 8.0 (6.0–10.8) | 1 | 1 | 1 |
| Yes | 7.3 (2.7–15.8) | 16.3 (8.8–30.3) | 4.8 (1.6–15.0) | 2.1 (1.0–4.5) | 1.1 (0.1–9.1) | 0.5 (0.1–2.4) |
| Antipsychotic use in the 5-year period before the index date | ||||||
| No | 5.3 (1.1–15.5) | 5.1 (2.5–10.1) | 1.6 (0.2–11.3) | 1 | 1 | 1 |
| Only atypical | 2.8 (1.6–4.6) | 20.1 (15.1–26.8) | 5.3 (3.2–8.9) | 0.5 (0.2–1.8) | 4.5 (2.1–9.7) | 2.8 (0.3–23.0) |
| Only typical | 12.6 (2.6–36.7) | 25.0 (14.8–42.2) | 13.3 (6.7–26.6) | 1.2 (0.2–6.1) | 3.8 (1.5–9.6) | 10.0 (1.1–92.8) |
| Typical and atypical | 3.4 (2.0–5.3) | 25.8 (19.4–34.3) | 11.0 (7.4–16.4) | 0.8 (0.2–2.6) | 4.1 (1.9–9.0) | 7.4 (0.9–59.3) |
| Antipsychotic use other than aripiprazole in 45 days before the index date | ||||||
| No b | 3.3 (1.9–5.3) | 17.4 (14.1–21.4) | 7.0 (4.6–10.5) | 1 | 1 | 1 |
| Only atypical | 2.6 (1.4–4.5) | 10.5 (5.2–20.9) | 8.6 (5.3–14.0) | 1.2 (0.6–2.3) | 0.5 (0.2–1.0) | 1.4 (0.7–2.7) |
| Only typical | 6.0 (1.6–15.5) | 35.0 (20.7–59.0) | 10.5 (5.2–21.0) | 1.8 (0.7–4.7) | 1.6 (0.8–3.0) | 1.0 (0.4–2.4) |
| Typical and atypical | 5.6 (2.3–11.6) | 265.4 (110.5–637.7) | 3.3 (0.5–23.6) | 2.0 (0.9–4.5) | 15.6 (5.3–45.6) | 0.3 (<0.1–2.4) |
EPS-related events were defined as pharmacy dispensings of anticholinergic drugs
Adjusted for the following variables: age; gender; number of previous diagnoses of EPS-related events in the 5 years before the index date; and diagnosis of depression, concomitant use of other psychotropics, use of antidepressants, and use of lithium in the 45 days prior to the index date.
Category includes users of oral aripiprazole and non-users of any antipsychotic.
CI, confidence interval; EPS, extrapyramidal symptoms.
Previous exposure to antipsychotics influenced the risk of EPS-related events during treatment with AOM. This was most evident in Italy, where prior use of antipsychotics during the 5-year period prior to the index date was associated with an increased risk for all exposure categories (i.e., for prior use of atypical antipsychotics, typical antipsychotics, and typical + atypical antipsychotics). In Sweden, IRs and the corresponding hazard ratios were also higher for patients with prior exposure to antipsychotics than for those without, with an increase in risk for previous users of typical antipsychotics only (Table 3). In Germany, there was no association between previous antipsychotic drug exposure and risk of EPS-related events, but this analysis was limited by low numbers of patients with events and broad confidence intervals. For the use of antipsychotic drugs (other than aripiprazole) in the 45 days before the index date, the influence was heterogeneous, with an increase in risk observed for users of typical and atypical antipsychotics in Italy.
Sensitivity analyses
Sensitivity analyses were conducted where the definition of an EPS-related event was based on diagnostic codes only (first and second sensitivity analysis), or where a combination of diagnostic codes and anticholinergic drug dispensings was required to identify an EPS-related event (third sensitivity analysis). All these analyses revealed substantially lower incidence rates than the main analysis (Table 4). This effect was most pronounced in Italy, where none of the patients under study had diagnostic codes of EPS during follow-up, resulting in incidence rates of zero for all sensitivity analyses.
Table 4.
Incidence rates of EPS-related events in sensitivity analyses.
| Approach (definition) | Incidence rates (95% CI) per 100 patient years | ||
|---|---|---|---|
| Germany | Italy | Sweden | |
| Main analysis (EPS-related events identified via anticholinergic drug dispensing) a | 3.4 (2.4–4.6) | 18.4 (15.3–22.1) | 7.7 (5.8–10.2) |
| First sensitivity analysis (EPS-related events identified via ICD-10 codes) b | 0.8 (0.4–1.5) | 0 | 0.3 (0.1–1.1) |
| Second sensitivity analysis (EPS-related events identified via ICD-10 codes, extended code list) b | 1.6 (1.0–2.5) | 0 | 1.2 (0.6–2.3) |
| Third sensitivity analysis (EPS-related events identified via ICD-10 codes, extended code list and required anticholinergic drug dispensing) a | 0.1 (<0.1–0.5) | 0 | 0.3 (0.1–1.2) |
Analysed in restricted study population.
Analysed in full study population (see Table S2 in Supplemental Material for patient numbers).
CI, confidence interval; EPS, extrapyramidal symptoms.
Discussion
In this multinational European health care database study, incidence rates of EPS-related events (defined as a first pharmacy dispensing of an anticholinergic drug) during treatment with AOM were lower in Germany and Sweden, as compared to Italy. The rates were highest during the first 30 days after initiation of treatment and decreased with longer duration of exposure.
A direct comparison of the observed study results with data from clinical trials is not possible due to the different ways of measuring the risks of EPS: While this observational database study used pharmacy dispensings of anticholinergic drugs as a proxy for EPS-related events, RCTs relied on adverse event reporting. If one assumes that all anticholinergic drug dispensings were made to treat actual symptoms of EPS (and not for prophylactic reasons), the cumulative incidences of anticholinergic drug dispensings could be expected to be lower or similar to the adverse event rates of EPS observed in clinical trials. The cumulative risk of EPS was 18.4% in the pooled pivotal studies ASPIRE (52 weeks) 8 and ASPIRE EU (38 weeks). 9 In this observational study, the cumulative incidences of EPS-related events, measured by first anticholinergic drug dispensings, were 10.0% in Germany (after 2 years), 11.5% in Sweden (after 2 years), 20.2% in Italy (after 1 year), and 27.8% in Italy (after 2 years). When considering the complexity of the outcome definition, these estimates are compatible with the EPS safety profile of AOM as characterized in clinical trials and included in the EU Summary of Product Characteristics. 12
In the main analysis, the definition of an EPS-related event is based on an incident pharmacy dispensing of an anticholinergic drug. Anticholinergic drugs may be used to treat EPS and thus may work as a proxy for the identification of EPS. However, the use of this proxy could lead to an overestimation or an underestimation of the actual EPS rate, depending on different methodological factors. Antipsychotic drugs prescribed concomitantly with AOM (i.e., during follow-up) may have contributed to the occurrence of EPS-related events, thereby increasing the EPS rate attributed to AOM exposure in this study. Prophylactic prescribing of anticholinergics for preventing EPS, although not recommended, 4 may have also caused an overestimation of the actual EPS rate. On the other hand, not all patients with EPS require treatment with anticholinergic drugs. According to a World Health Organization (WHO) recommendation, 4 short-term use of anticholinergics should be restricted to individuals with significant extrapyramidal side effects when dose reduction and switching strategies have proven ineffective, or when these side effects are acute or severe. This may have led to an underestimation of the actual EPS rate, especially in the case of less severe EPS symptoms. In addition, this aspect underlines that EPS adverse event reporting frequencies in RCTs can only be used for contextualization, but not for direct comparison with the results of this study.
Estimates of incidence rates and cumulative incidences of EPS-related events were considerably higher in Italy than in Germany or Sweden. Considering that anticholinergic drug dispensings were used to define EPS-related events, regional differences in the incidence rates of EPS-related events may reflect differences in prescribing habits rather than differences in symptom frequencies. Data from previous observational studies indicate that there is large regional variation in prescribing anticholinergic drugs, with high prescription rates in Southern Europe including Italy: The European part of the prospective Schizophrenia Outpatients Health Outcomes (SOHO) study showed that baseline use of anticholinergic drugs was higher in Southern Europe (42.6% of patients) than in Central/Eastern Europe (33.1%) or Northern Europe (26.2%). 19 In contrast, the proportion of patients with recorded EPS at baseline was similar in the three regions (Southern Europe: 50.1%; Central/Eastern Europe: 49.4%; Northern Europe: 45.8%). The large variation in concomitant medication use (including anticholinergics) among the countries participating in SOHO persisted, even after adjusting for the type of antipsychotic drug used and patients’ clinical characteristics. 21 This indicates that national or regional prescribing habits importantly influenced anticholinergic drug prescribing at the time this study was conducted. It seems reasonable to assume that at least some of these differences persisted until more recent time periods. The SOHO study also revealed that patients from Italy had a 50% higher likelihood of receiving anticholinergic drugs than patients from Germany or Denmark, even after adjusting for several patient and treatment characteristics. 21 This is in line with the findings of this PASS. An Italian retrospective cohort study also found that use of anticholinergics was high in patients exposed to first-generation depot antipsychotics during 2 years of treatment (46.7%). 22 Substantial regional variation in anticholinergic drug prescribing rates was also reported for other regions. Xiang et al. 23 reported on such variations in Asia, where prescribing of anticholinergic drugs was 3.3-fold higher in Japan than in China. Pristed et al. 24 reported for Denmark that prescription patterns varied considerably between national regions, ranging from 4.0% in the Capital Region to 8.1% in the Northern Denmark Region.
The Italian database only covered two regions in Southern Italy (Palermo and Caserta), while the German and Swedish databases were representative of the overall country. As it is unknown to what extent the anticholinergic drug prescribing habits in Southern Italy are representative of the country in total, this may have aggravated the resulting differences between Italy and Germany/Sweden. Differences in terms of health service utilization were observed at baseline (e.g., lower previous hospitalisation rates in Italy), which may have also contributed to differences in outpatient anticholinergic drug prescribing.
Previous exposure to antipsychotics had an influence on anticholinergic drug prescribing during AOM treatment, especially in Italy. For patients without previous use of antipsychotics, the incidence rates of EPS-related events were low in all three countries, indicating that the differences between countries were mainly caused by patients with an antipsychotic drug history. This may, for instance, be due to patients’ experiences with antipsychotics, EPS, and treatment of EPS, leading to proactively requesting anticholinergic drugs from their physicians. In addition, a cumulative effect of previous exposure, or physicians’ experiences, preferences and expectations might have also played a role. The low rates for those without previous use of antipsychotics in all three countries indicate that the decision to prescribe anticholinergic drugs is not only influenced by the actual pharmacological effect of AOM, but also by other factors in the individual history of each patient.
Limitations
Some limitations have to be considered when interpreting the results of this PASS: The study did not include a control group, and thus, results from RCTs were used as an external reference to contextualize the study findings of cumulative incidences of EPS-related events. Drug exposure information in this study were based on pharmacy dispensings included in the respective databases. Information on prescribed daily doses of drugs or the intended way of anticholinergic drug use (as-needed, regular intake) were not available. Also, the databases did not include information on in-hospital drug treatment with antipsychotics (including AOM) or anticholinergic drugs. Using anticholinergic drug dispensings as a proxy for the risk of EPS represents a methodological limitation of this study. However, the number of patients using antiparkinsonian drugs has also been used as a proxy for extrapyramidal symptoms in previous research. A network meta-analysis of randomized clinical trials used this definition to compare side effects of oral versus long-acting injectable antipsychotics. 25 In addition, the sensitivity analyses clearly indicated that using EPS diagnoses instead of anticholinergic drug dispensings to define the primary outcome would have led to a substantial underestimation of the true event rates. The unavailability of diagnoses from primary care in the Swedish and Italian databases might have contributed to these findings, but the incidence rates based on diagnoses were also low in Germany, where diagnoses from primary care were available. This could reflect that physicians may rarely use ICD codes to indicate the occurrence of EPS. The analysis of the proportion of patients with recorded use of oral aripiprazole before initiation of AOM was limited by the fact that drugs administered during a hospital stay that had preceded the initiation of AOM were not available in the databases. In addition, ambulatory dispensings of oral aripiprazole would have been missed if these occurred more than 45 days prior to AOM initiation. The actual use of oral aripiprazole before initiation of AOM may have thus been underestimated. The same applies to analyses in relation to the use of antipsychotic drugs other than aripiprazole. If AOM was initiated during an in-hospital stay, the study would have identified the first ambulatory AOM dispensing after hospital discharge as the first exposure and as the start of follow-up (index date). This may have influenced cumulative incidence estimates. Diabetes mellitus and previous exposure to antipsychotic drugs were identified as potential risk factors for the occurrence of EPS-related events. However, as these results were not consistent across all countries and the CIs of the hazard ratios were wide, this finding needs to be interpreted with caution. It was not possible to separate prophylactic from therapeutic dispensings of anticholinergic drugs. This was addressed in part by not considering dispensings of anticholinergic drugs that occurred already on the index date or the day thereafter, but it may still have contributed to an overestimation of the actual rate of EPS. The study was not designed to identify causes for the observed regional differences in anticholinergic drug prescribing or dispensing in patients treated with AOM.
Conclusion
In this multinational, European health care real-world database study, the cumulative incidences of EPS-related events, defined as the first dispensing of an anticholinergic drug during follow-up, were compatible with the known safety profile of AOM. IRs of EPS-related events were lower in Germany or Sweden than in Italy. The regional variation is most likely attributable to different prescribing habits for anticholinergic drugs, which is in line with previous findings. Additional studies would be required to verify this interpretation.
Supplemental Material
Supplemental material, sj-docx-1-tpp-10.1177_20451253251368010 for Frequency of anticholinergic drug use in patients treated with aripiprazole once-monthly: a 2-year cohort study using European healthcare databases by Frank Andersohn, Pedro Such, Michael Jan, Uwa Kalu, Jessica McDonough, Jonas Reinold, Oliver Riedel, Gianluca Trifirò, Valentina Ientile, Michele Tari, Maurizio Pastorello, Alejandro Arana, Joan Forns, Katja M. Hakkarainen, Leyla Nunez, Kristian Tore Jørgensen, Jacob Simonsen and Murat Yildirim in Therapeutic Advances in Psychopharmacology
Supplemental material, sj-docx-2-tpp-10.1177_20451253251368010 for Frequency of anticholinergic drug use in patients treated with aripiprazole once-monthly: a 2-year cohort study using European healthcare databases by Frank Andersohn, Pedro Such, Michael Jan, Uwa Kalu, Jessica McDonough, Jonas Reinold, Oliver Riedel, Gianluca Trifirò, Valentina Ientile, Michele Tari, Maurizio Pastorello, Alejandro Arana, Joan Forns, Katja M. Hakkarainen, Leyla Nunez, Kristian Tore Jørgensen, Jacob Simonsen and Murat Yildirim in Therapeutic Advances in Psychopharmacology
Acknowledgments
For Germany: We thank all statutory health insurance providers that provided data for this study, namely AOK Bremen/Bremerhaven, DAK-Gesundheit, Die Techniker (TK), and hkk Krankenkasse, as well as Alina Ludewig and Inga Schaffer for programming the datasets and tables.
Footnotes
ORCID iDs: Frank Andersohn 
https://orcid.org/0000-0003-3514-9238
Jonas Reinold
https://orcid.org/0000-0001-8266-2574
Supplemental material: Supplemental Material for this article is available online.
Contributor Information
Frank Andersohn, Frank Andersohn Consulting & Research Services, Mandelstr 16, Berlin 10409, Germany.
Pedro Such, H. Lundbeck A/S, Valby, Denmark.
Michael Jan, Otsuka Pharmaceutical Development & Commercialization, Inc., Princeton, NJ, USA.
Uwa Kalu, Lundbeck, Deerfield, IL, USA.
Jessica McDonough, Otsuka Pharmaceutical Development & Commercialization, Inc., Princeton, NJ, USA.
Jonas Reinold, Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany.
Oliver Riedel, Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany.
Gianluca Trifirò, Department of Diagnostics and Public Health, University of Verona, Verona, Italy.
Valentina Ientile, Department of Biomedical and Dental Sciences and Morphofunctional Imaging, University of Messina, Messina, Italy.
Michele Tari, Caserta Local Health Unit, Caserta, Italy.
Maurizio Pastorello, Department of Pharmacy, Local Health Unit of Palermo, Palermo, Italy.
Alejandro Arana, RTI Health Solutions, Barcelona, Spain.
Joan Forns, RTI Health Solutions, Barcelona, Spain.
Katja M. Hakkarainen, IQVIA Global Database Studies, Mölndal, Sweden Epidemiology & Real-World Science, RWE Scientific Affairs, Parexel International, Gothenburg, Sweden.
Leyla Nunez, IQVIA Global Database Studies, Mölndal, Sweden.
Kristian Tore Jørgensen, H. Lundbeck A/S, Valby, Denmark.
Jacob Simonsen, H. Lundbeck A/S, Valby, Denmark.
Murat Yildirim, H. Lundbeck A/S, Valby, Denmark.
Declarations
Ethics approval and consent to participate: This article is based on previously collected human health data. No human participants were recruited for this study. The researchers of each database followed the respective regulations for internal ethics and scientific study approval, for example, review and approval by the institutions providing data access. For the German database, all involved health insurance providers, as well as the German Federal Office for Social Security and the Senator for Health, Women and Consumer Protection in Bremen, as their responsible authorities, approved the use of GePaRD data for this study. For the Swedish database (national registries), the study received ethics committee approval from the Swedish Ethical Review Authority, as well as data permits from the National Board of Health and Welfare (Socialstyrelsen) and Statistics Sweden (Statistiska centralbyrån). For the database in Italy, the study protocol was approved by the Ethical Committee of the Academic Hospital of Messina (N. prot.08/20), which had at that time in place agreements with both Local Health Units (LHUs) of Palermo and Caserta for access to anonymized claims data from those LHUs.
Consent for publication: Not applicable.
Author contributions: Frank Andersohn: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Software; Visualization; Writing – original draft; Writing – review & editing.
Pedro Such: Conceptualization; Methodology; Supervision; Writing – review & editing.
Michael Jan: Conceptualization; Methodology; Supervision; Writing – review & editing.
Uwa Kalu: Conceptualization; Methodology; Supervision; Writing – review & editing.
Jessica McDonough: Methodology; Project administration; Writing – review & editing.
Jonas Reinold: Formal analysis; Methodology; Supervision; Writing – review & editing.
Oliver Riedel: Formal analysis; Methodology; Writing – review & editing.
Gianluca Trifirò: Formal analysis; Methodology; Writing – review & editing.
Valentina Ientile: Formal analysis; Methodology; Writing – review & editing.
Michele Tari: Formal analysis; Methodology; Writing – review & editing.
Maurizio Pastorello: Formal analysis; Methodology; Writing – review & editing.
Alejandro Arana: Formal analysis; Methodology; Writing – review & editing.
Joan Forns: Formal analysis; Methodology; Writing – review & editing.
Katja M. Hakkarainen: Formal analysis; Methodology; Writing – review & editing.
Leyla Nunez: Formal analysis; Methodology; Writing – review & editing.
Kristian Tore Jørgensen: Conceptualization; Methodology; Writing – review & editing.
Jacob Simonsen: Conceptualization; Formal analysis; Methodology; Writing – review & editing.
Murat Yildirim: Conceptualization; Methodology; Supervision; Writing – review & editing.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was sponsored by Otsuka Pharmaceutical Development & Commercialization Inc. (Princeton, NJ, USA) and H. Lundbeck A/S (Valby, Denmark). The sponsors were involved in the design of the study, the collection, analysis and interpretation of data, the writing and reviewing of this article, and the decision to submit the article for publication. Database analyses, reporting and interpretation of results were obligations of the respective database research teams and were not influenced by the study sponsor.
The co-authors declare the following competing interests: JS, KTJ, MY, PS, and UK are/were employees at Lundbeck, the manufacturer of aripiprazole once monthly. MJ and JM were employees of Otsuka and have collaborated with Lundbeck. KMH and LN are/were employees of IQVIA, a company that conducts commissioned pharmacoepidemiological studies for various pharmaceutical companies. AA and JF are employees at RTI-HS, an institute contracted by Lundbeck. OR and JR are working at an independent, non-profit research institute, the Leibniz Institute for Prevention Research and Epidemiology—BIPS. This post-authorization safety study (PASS) was requested by the EMA, which required that Lundbeck finance the study. The BIPS authors had complete autonomy in the process of establishing the protocol, carrying out the analyses, and interpreting the results. In addition, GT has received payments for lectures, presentations, and advisory boards from several pharmaceutical companies, FA has received consulting fees and honoraria from several pharmaceutical companies, including one of the study sponsor companies (Lundbeck), KMH is an employee of Parexel International, which conducts contracted research for several pharmaceutical companies. VL, MP, MT have not reported conflicts of interest.
Availability of data and materials: Access to patient-level data is only possible for members of the respective national study team due to national data protection regulations.
References
- 1. Ali T, Sisay M, Tariku M, et al. Antipsychotic-induced extrapyramidal side effects: a systematic review and meta-analysis of observational studies. PLoS One 2021; 16: e0257129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Barnes TR, McPhillips MA. Novel antipsychotics, extrapyramidal side effects and tardive dyskinesia. Int Clin Psychopharmacol 1998; 13(Suppl. 3): 49–57. [DOI] [PubMed] [Google Scholar]
- 3. Divac N, Prostran M, Jakovcevski I, et al. Second-generation antipsychotics and extrapyramidal adverse effects. Biomed Res Int 2014; 2014: 656370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. WHO. Role of anticholinergic medications in patients requiring long-term antipsychotic treatment for psychotic disorders, https://cdn.who.int/media/docs/default-source/mental-health/mhgap/psychosis-and-bipolar-disorders/role-of-anticholinergic-medications-in-patients-requiring-long-term-antipsychotic-treatment-for-psychotic-disorders.pdf?sfvrsn=68417542_0 (2012, accessed 29 September 2024).
- 5. Ribeiro ELA, de Mendonca Lima T, Vieira MEB, et al. Efficacy and safety of aripiprazole for the treatment of schizophrenia: an overview of systematic reviews. Eur J Clin Pharmacol 2018; 74: 1215–1233. [DOI] [PubMed] [Google Scholar]
- 6. Fusar-Poli P, Kempton MJ, Rosenheck RA. Efficacy and safety of second-generation long-acting injections in schizophrenia: a meta-analysis of randomized-controlled trials. Int Clin Psychopharmacol 2013; 28: 57–66. [DOI] [PubMed] [Google Scholar]
- 7. Park SC, Choi MY, Choi J, et al. Comparative efficacy and safety of long-acting injectable and oral second-generation antipsychotics for the treatment of schizophrenia: a systematic review and meta-analysis. Clin Psychopharmacol Neurosci 2018; 16: 361–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kane JM, Sanchez R, Perry PP, et al. Aripiprazole intramuscular depot as maintenance treatment in patients with schizophrenia: a 52-week, multicenter, randomized, double-blind, placebo-controlled study. J Clin Psychiatry 2012; 73: 617–624. [DOI] [PubMed] [Google Scholar]
- 9. Fleischhacker WW, Sanchez R, Perry PP, et al. Aripiprazole once-monthly for treatment of schizophrenia: double-blind, randomised, non-inferiority study. Br J Psychiatry 2014; 205: 135–144. [DOI] [PubMed] [Google Scholar]
- 10. Etminan M, Procyshyn RM, Samii A, et al. Risk of extrapyramidal adverse events with aripiprazole. J Clin Psychopharmacol 2016; 36: 472–474. [DOI] [PubMed] [Google Scholar]
- 11. von Elm E, Altman DG, Egger M, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577. [DOI] [PubMed] [Google Scholar]
- 12. Otsuka Pharmaceutical Netherlands B.V. Abilify Maintena Summary of Product Characteristics, https://www.ema.europa.eu/en/documents/product-information/abilify-maintena-epar-product-information_en.pdf (2008, accessed 29 September 2024).
- 13. Pierre MJ. Extrapyramidal symptoms with atypical antipsychotics incidence, prevention and management. Drug Saf 2005; 28: 191–208. [DOI] [PubMed] [Google Scholar]
- 14. Tenback DE, van Harten PN. Epidemiology and risk factors for (tardive) dyskinesia. Int Rev Neurobiol 2011; 98: 211–230. [DOI] [PubMed] [Google Scholar]
- 15. Woerner MG, Alvir JM, Saltz BL, et al. Prospective study of tardive dyskinesia in the elderly: rates and risk factors. Am J Psychiatry 1998; 155: 1521–1528. [DOI] [PubMed] [Google Scholar]
- 16. Woods SW, Morgenstern H, Saksa JR, et al. Incidence of tardive dyskinesia with atypical versus conventional antipsychotic medications: a prospective cohort study. J Clin Psychiatry 2010; 71: 463–474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Weiden PJ. EPS profiles: the atypical antipsychotics are not all the same. J Psychiatr Pract 2007; 13: 13–24. [DOI] [PubMed] [Google Scholar]
- 18. Glazer WM. Review of incidence studies of tardive dyskinesia associated with typical antipsychotics. J Clin Psychiatry 2000; 61(Suppl. 4): 15–20. [PubMed] [Google Scholar]
- 19. Karagianis J, Novick D, Pecenak J, et al. Worldwide-Schizophrenia Outpatient Health Outcomes (W-SOHO): baseline characteristics of pan-regional observational data from more than 17,000 patients. Int J Clin Pract 2009; 63: 1578–1588. [DOI] [PubMed] [Google Scholar]
- 20. Aalen OO, Johansen S. An empirical transition matrix for non-homogeneous Markov Chains based on censored observations. Scandinavian J Stat 1978; 5: 141–150. [Google Scholar]
- 21. Novick D, Bousono M, Suarez D, et al. Use of concomitant medication with antipsychotic treatment in outpatients with schizophrenia: results from the European Schizophrenia Outpatients Health Outcomes (SOHO) study. Prog Neuropsychopharmacol Biol Psychiatry 2005; 29: 972–982. [DOI] [PubMed] [Google Scholar]
- 22. Marchiaro L, Rocca P, LeNoci F, et al. Naturalistic, retrospective comparison between second-generation antipsychotics and depot neuroleptics in patients affected by schizophrenia. J Clin Psychiatry 2005; 66: 1423–1431. [DOI] [PubMed] [Google Scholar]
- 23. Xiang YT, Ungvari GS, Correll CU, et al. Trends in the access to and the use of antipsychotic medications and psychotropic co-treatments in Asian patients with schizophrenia. Epidemiol Psychiatr Sci 2016; 25: 9–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Pristed SG, Correll CU, Nielsen J. Frequency and correlates of anticholinergic use among patients with schizophrenia in Denmark: a Nation-wide pharmacoepidemiological study. Psychiatry Res 2017; 255: 198–203. [DOI] [PubMed] [Google Scholar]
- 25. Wang D, Schneider-Thoma J, Siafis S, et al. Efficacy, acceptability and side-effects of oral versus long-acting- injectables antipsychotics: systematic review and network meta-analysis. Eur Neuropsychopharmacol 2024; 83: 11–18. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, sj-docx-1-tpp-10.1177_20451253251368010 for Frequency of anticholinergic drug use in patients treated with aripiprazole once-monthly: a 2-year cohort study using European healthcare databases by Frank Andersohn, Pedro Such, Michael Jan, Uwa Kalu, Jessica McDonough, Jonas Reinold, Oliver Riedel, Gianluca Trifirò, Valentina Ientile, Michele Tari, Maurizio Pastorello, Alejandro Arana, Joan Forns, Katja M. Hakkarainen, Leyla Nunez, Kristian Tore Jørgensen, Jacob Simonsen and Murat Yildirim in Therapeutic Advances in Psychopharmacology
Supplemental material, sj-docx-2-tpp-10.1177_20451253251368010 for Frequency of anticholinergic drug use in patients treated with aripiprazole once-monthly: a 2-year cohort study using European healthcare databases by Frank Andersohn, Pedro Such, Michael Jan, Uwa Kalu, Jessica McDonough, Jonas Reinold, Oliver Riedel, Gianluca Trifirò, Valentina Ientile, Michele Tari, Maurizio Pastorello, Alejandro Arana, Joan Forns, Katja M. Hakkarainen, Leyla Nunez, Kristian Tore Jørgensen, Jacob Simonsen and Murat Yildirim in Therapeutic Advances in Psychopharmacology