Journal of Affective Disorders 249 (2019) 242–252 Contents lists available at ScienceDirect Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad Research paper Antidepressant use in Denmark, Germany, Spain, and Sweden between 2009 T and 2014: Incidence and comorbidities of antidepressant initiators ⁎ Joan Fornsa, , Anton Pottegårdb, Tammo Reindersc, Beatriz Poblador-Ploud, Rosa Morrose,f,g, Lena Brandth, Miguel Cainzos-Achiricaa, Maja Hellfritzschb, Tania Schinkc, Alexandra Prados-Torresd, Maria Giner-Sorianoe,f,g, David Häggh, Jesper Hallasb, Jordi Cortése,f,g,i, Emmanuelle Jacquotj, Nicolas Deltourj, Susana Perez-Gutthanna, Manel Pladevalla, Johan Reutforsh a Epidemiology, RTI Health Solutions, Barcelona, Spain b Clinical Pharmacology and Pharmacy, Department of Public Health, University of Southern Denmark, Odense, Denmark c Leibniz Institute for Prevention Research and Epidemiology–BIPS, Bremen, Germany d EpiChron Research Group on Chronic Diseases, Aragon Health Sciences Institute (IACS), IIS Aragón, REDISSEC ISCIII, Miguel Servet University Hospital, Zaragoza, Spain e Fundació Institut Universitari per a la recerca a l'Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol), Barcelona, Spain f Universitat Autònoma de Barcelona, Bellaterra (Cerdanyola del Vallès), Spain g Institut Català de la Salut, Barcelona, Spain h Centre for Pharmacoepidemiology, Department of Medicine, Solna, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden i Universitat Politècnica de Catalunya, Departament d'Estadística i Investigació Operativa, Barcelona, Spain j Pharmacoepidemiology Department, Les Laboratoires Servier, Suresnes, France A R T I C LE I N FO A B S T R A C T Keywords: Background: We aimed to describe patterns of use and characteristics of 10 commonly used antidepressants for Antidepressants the period 2009–2014 in Denmark, Germany, Spain, and Sweden. Comorbidity Methods: Adult initiators from 2009 to 2014 of each study antidepressant were identified in four countries using Drug utilization study five data sources: the Danish National registers, GePaRD (Germany), EpiChron (Aragon, Spain), SIDIAP Europe (Catalonia, Spain), and the Swedish National Registers. The study included 10 study antidepressants: citalopram, escitalopram, fluoxetine, paroxetine, sertraline, duloxetine, venlafaxine, amitriptyline, mirtazapine, and ago- melatine. Results: Citalopram was the most prescribed study antidepressant, followed by mirtazapine. Paroxetine and agomelatine were the least prescribed antidepressants. Mirtazapine was widely used among older antidepressant initiators with higher percentages of comorbidities at baseline, and fluoxetine was used among young patients. Citalopram and amitriptyline had the lowest percentage of multiple antidepressant use in the 12 months prior to the current treatment episode, while agomelatine, duloxetine, and venlafaxine had the highest percentage of multiple antidepressant use in the year prior to the current treatment episode. Limitations: The most important limitations are exposure information based on filled prescriptions, focus on antidepressant initiators only, lack of information on the indication, and heterogeneity of the type of data across data sources. Conclusions: Results of this study including 4.8 million study antidepressant initiators of study antidepressants suggest that citalopram and mirtazapine are the most commonly prescribed antidepressants. Agomelatine and paroxetine were the least used antidepressants in the participating populations. Mirtazapine was the anti- depressant most commonly prescribed among older antidepressant initiators with high percentage of co- morbidities at baseline, whereas fluoxetine was commonly used among young patients. ⁎ Corresponding author. E-mail address: jforns@rti.org (J. Forns). https://doi.org/10.1016/j.jad.2019.02.010 Received 11 October 2018; Received in revised form 18 January 2019; Accepted 5 February 2019 Available online 06 February 2019 0165-0327/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 1. Introduction antidepressant use was an exclusion criterion. Antidepressants are among the most prescribed drugs in Europe. Besides depression as the major indication, antidepressants are also 2.3. Study antidepressants used for a wide range of other conditions such as panic disorders, generalized anxiety disorder, and neuropathic pain (Noordam et al., The selected study antidepressants included five SSRIs (citalopram, 2015). The most recent drug utilization studies conducted in Europe escitalopram, fluoxetine, paroxetine, and sertraline), two serotonin- suggest an increase in the use of antidepressants over time, particularly norepinephrine reuptake inhibitors (SNRIs) (duloxetine and venla- for selective serotonin reuptake inhibitors (SSRIs), the most commonly faxine), one tricyclic antidepressant (TCA) (amitriptyline), one nora- prescribed type of antidepressant class (Abbing-Karahagopian et al., drenaline and specific serotonergic antidepressants (NASSAs) (mirta- 2014; Bauer et al., 2008; Noordam et al., 2015; Poluzzi et al., 2013). zapine), and one melatonergic agonist and 5-HT2C antagonist In past decades, the number of antidepressants available on the (agomelatine) (Supplementary Table 2). These antidepressants were market has increased markedly, resulting in a great variety of treatment selected in the agomelatine postauthorisation safety study to represent options for prescribers. The choice of an antidepressant is influenced by (1) the most commonly used antidepressants in the participating several factors, including those specifically related to the drug profile countries and (2) different classes of antidepressants. Agomelatine was (e.g. side effects, tolerability, and cost), physician characteristics (e.g. selected for regulatory reasons as the main exposure of interest specialty, country of practice), reimbursement policies (Bauer et al., (Pladevall, 2018). Specific details on the exposure definition are in- 2008), and patient characteristics, such as severity of depression and cluded in Supplementary Methods. Indication of use of study anti- presence of comorbidities (Jobski et al., 2017b; Zimmerman et al., depressants for the study population was not available. 2004). With the exception of one German study restricted to adults older 2.4. Characterisation of initiators of study antidepressants than 65 years (Jobski et al., 2017b), no studies have analysed the use of antidepressants in adults from the general population in several Eur- Initiators of the study antidepressants were characterised at the date opean countries since 2012. Also, little information is available on the of their first prescription according to age and sex. Comorbidity in- patterns of comorbidities among initiators of different antidepressant formation at any time before cohort inclusion was ascertained through drugs. Therefore, we aimed to describe the cumulative incidence of hospital discharge diagnoses and diagnoses reported in primary care, initiators of 10 commonly prescribed antidepressants for the period when available, in combination with information on the use of medi- 2009–2014 in Denmark, Germany, Spain, and Sweden. We also aimed cations for specific diseases when applicable (Pladevall, 2018). In- to compare the characteristics of initiators of each of the 10 study an- formation on hospitalisations during the 12 months prior to the current tidepressants in terms of their demographic characteristics, presence of treatment episode was also collected as a measure of health care utili- comorbidities, health care resource use, and patterns of use. sation. Patterns of antidepressant use were also assessed. First, we mea- 2. Methods sured the recent use of any antidepressant (N06A group) available on the market at the time of the study (0, 1, 2, or more) in the 12 months 2.1. Study setting prior to the current treatment episode. Switching and multiple uses were not differentiated but rather combined into a single variable. We This drug utilization study was conducted using data collected also measured the duration of the first treatment episode of current use during a postauthorisation safety study investigating the potential risk (in months). Finally, we measured whether the study antidepressants of acute hepatoxic reactions associated with the use of agomelatine and were prescribed alone or combined with other antidepressants by cal- other antidepressants (Pladevall, 2018). Initiators of each study anti- culating the percentage of treatment episodes of current use with depressant were identified in four countries using five automated health combined use of another antidepressant. data sources: the Danish National Health Registers (Denmark) (Pottegard et al., 2016; Schmidt et al., 2015), the German Pharma- coepidemiological Research Database (GePaRD) (Germany) (Jobski 2.5. Statistical analyses et al., 2017a,b; Pigeot and Ahrens, 2008), the EpiChron Cohort from Aragon Health Sciences Institute (Aragon, Spain) (Prados-Torres et al., Data describing the study population and prescribing patterns of 2018), the Information System for Research in Primary Care (SIDIAP) antidepressants are presented as means, standard deviations, fre- (Catalonia, Spain) (SIDIAP, 2014), and the Swedish National Registers quencies, or percentages, as appropriate. The estimated cumulative (Sweden) (Ludvigsson et al., 2011; Wettermark et al., 2007). The main incidence of antidepressant initiation at the end of the study period was characteristics of each database are described in Supplementary calculated overall and for each antidepressant by dividing the total Table 1. The study period in each data source started after the launch of number of initiators and the number of initiators for each study anti- agomelatine in each respective country (2009 or 2010) and ended on depressant during the study period (2009–2013/2014) by the total re- the last year for which data was available in each data source (2013 or ference adult population in each data source at the end of 2013 or 2014 2014). (depending on the end of the study period). Stata (version 14) or SAS (version 9.4 or later) was used for the analyses by the different study 2.2. Study population research partners evaluating each of the individual data sources. The study cohort included all adult initiators of any of the 10 re- levant antidepressants in the study databases from 2009 to 2014 with at 3. Results least 12 months of continuous enrolment in the data source. The study cohort included all individuals aged 18 years or older at the date of the The study included a total of 4,833,774 initiators of antidepressants first-recorded prescription fill for any of the study antidepressants (Table 1). Sweden had the largest population with 1.8 million initiators during the study period(s) with no prescription fill for the same study of study antidepressants, followed by GePaRD (Germany) with 1.7 antidepressant within the prior 12 months (initiators). One patient million, Denmark with 0.8 million, and the Spanish populations (Epi- could contribute to several antidepressant groups if eligibility criteria Chron [Aragon] and SIDIAP [Catalonia]) with less than 0.3 million were accomplished. For women, pregnancy at the start date of each. 243 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 Table 1 Number of study antidepressant initiators with age and sex distribution. Denmark GePaRD, Germany EpiChron, Aragon, Spain SIDIAP, Catalonia, Spain Sweden Study period Number of study antidepressant initiators, n (%) Citalopram 253,531 (30%) 463,794 (27%) 10,412 (5%) 53,076 (20%) 405,944 (22%) Escitalopram 57,184 (7%) 77,765 (5%) 52,256 (25%) 29,429 (11%) 141,701 (8%) Fluoxetine 18,448 (2%) 66,782 (4%) 17,045 (8%) 23,784 (9%) 77,044 (4%) Paroxetine 18,968 (2%) 50,141 (3%) 26,213 (12%) 38,883 (15%) 34,666 (2%) Sertraline 137,644 (16%) 88,200 (5%) 12,899 (6%) 28,101 (11%) 359,636 (20%) Duloxetine 37,514 (4%) 79,456 (5%) 23,744 (11%) 13,509 (5%) 86,454 (5%) Venlafaxine 87,390 (10%) 147,580 (9%) 11,951 (6%) 15,577 (6%) 130,716 (7%) Amitriptyline 39,404 (5%) 309,377 (18%) 22,716 (11%) 33,870 (13%) 204,078 (11%) Mirtazapine 167,020 (20%) 370,741 (22%) 23,971 (11%) 20,324 (8%) 347,596 (19%) Agomelatine 21,941 (3%) 62,009 (4%) 10,077 (5%) 4167 (2%) 19,046 (1%) Total 839,044 1,715,845 211,284 260,720 1,806,881 Women (%) Citalopram 61% 67% 69.2% 70.4% 65.3% Escitalopram 61% 66.7% 69.1% 67.1% 64.9% Fluoxetine 69.9% 72.2% 75.2% 75% 73.5% Paroxetine 60.5% 65.7% 71.5% 70.3% 60.5% Sertraline 62.4% 67% 67.2% 69.8% 63.6% Duloxetine 65.1% 69.3% 73.7% 75.1% 65.6% Venlafaxine 61.1% 65.9% 69.5% 70.2% 60.6% Amitriptyline 65.3% 71.8% 74.6% 75.7% 69% Mirtazapine 57.5% 65.9% 65.5% 64.7% 58.5% Agomelatine 65% 68.3% 69.7% 67.4% 62.1% Age, median (IQR), years Citalopram 51 (36–70) 54 (42–70) 63 (46–79) 55 (41–71) 56 (39–75) Escitalopram 52 (37–69) 51 (39–63) 56 (42–73) 51 (38–67) 44 (32–59) Fluoxetine 41 (29–54) 47 (36–57) 50 (39–63) 47 (37–59) 38 (26–51) Paroxetine 44 (34–59) 49 (37–60) 53 (40–68) 49 (37–63) 45 (32–58) Sertraline 44 (31–59) 52 (41–65) 62 (44–78) 61 (44–76) 43 (30–60) Duloxetine 48 (37–60) 57 (47–70) 59 (46–73) 56 (45–68) 48 (37–61) Venlafaxine 45 (33–57) 51 (40–61) 55 (43–70) 53 (42–67) 44 (32–58) Amitriptyline 51 (41–65) 58 (47–71) 53 (40–67) 52 (39–65) 55 (43–68) Mirtazapine 54 (40–72) 58 (47–73) 67 (49–81) 60 (45–76) 58 (41–76) Agomelatine 46 (36–57) 52 (42–61) 54 (42–67) 51 (41–62) 45 (33–56) IQR = interquartile range. 3.1. Demographic characteristics of antidepressant initiators Table 2 Cumulative incidence of study antidepressant initiation at the end of study Women comprised the majority of initiators for all antidepressants period (per 1000 population). (Table 1). The study antidepressant with the highest proportion of Denmarka GePaRD, EpiChron, SIDIAP, Swedena women initiators was fluoxetine (>70% in four of five populations). In Germanyb Aragon, Catalonia, general, fluoxetine was prescribed to younger patients (median age Spainc Spaina range, 38 years in Sweden to 50 years in Aragon [Spain]), and mirta- Citalopram 64.5 43.7 9.5 38.0 52.5 zapine was consistently prescribed to older patients (median range, 54 Escitalopram 14.5 7.3 47.5 21.1 18.3 years in Denmark to 67 years in Aragon [Spain]). Among SSRIs, cita- Fluoxetine 4.7 6.3 15.5 17.0 10.0 lopram was commonly prescribed among older patients (range, 51 Paroxetine 4.8 4.7 23.8 27.8 4.5 years in Denmark to 63 years in Aragon [Spain]). Sertraline 35.0 8.3 11.7 20.1 46.5 Duloxetine 9.5 7.5 21.6 9.7 11.2 Venlafaxine 22.2 13.9 10.9 11.1 16.9 Amitriptyline 10.0 29.2 20.6 24.2 26.4 3.2. Incidence of antidepressant initiation Mirtazapine 42.5 35.0 21.8 14.5 45.0 Agomelatine 5.6 5.9 9.2 3.0 2.5 The highest total cumulative incidence of antidepressant initiation Total 213.3 161.8 192.0 186.5 233.8 for the 10 antidepressants (as per 1000 population) during the period 2009 and 2013/2014 (depending on data source) was found in Sweden Note: The estimated cumulative incidence of study antidepressant initiation at the end of the study period was calculated by dividing the total number of with 234, followed by Denmark, 213; Aragon (Spain), 192; Catalonia initiators during the study period (2009–2013/2014) of each antidepressant by (Spain), 187; and Germany, 162 (Table 2). Overall, citalopram was the the total reference adult population in each data source at the end of 2013 or most prescribed study antidepressant, except in Aragon (Spain), with a 2014 (depending on the end of the study period). cumulative incidence ranging from 10 per 1000 population in Aragon a Population reference: adult population in the data source at 01 January (Spain) to 65 per 1000 population in Denmark. In Aragon (Spain), the 2014. most prescribed study antidepressant was escitalopram, with a cumu- b Population reference: average of the adult population average between the lative incidence of 48 per 1000 population. For the other SSRIs, we size of the source population at the start of the study period (2009) and the end observed higher rates for sertraline in Denmark and Sweden when of the study period (2013). c compared with other populations. In contrast, rates for fluoxetine and Population reference: adult population in the data source at 01 January paroxetine were higher in Aragon and Catalonia populations than in the 2013. other populations. The use of duloxetine and venlafaxine was similar in all populations, with a higher cumulative incidence for venlafaxine when compared with that for duloxetine, except in Aragon (Spain). 244 Table 3 Comorbid conditions among study antidepressant initiators. SSRIs,% (95% CI) SNRIs,% (95% CI) TCA,% (95% CI) NASSA,% (95% Other,% (95% J. Forns, et al. CI) CI) Citalopram Escitalopram Fluoxetine Paroxetine Sertraline Duloxetine Venlafaxine Amitriptyline Mirtazapine Agomelatine Overall Denmark (n = 839,044) Hypertension 33.3 32.8 (31.6–34.0) 20 (19.0–21.0) 25.1 25.1 30.2 24.5 37.6 (36.3–38.9) 36.6 (35.3–37.9) 27.0 (25.9–28.1) 31.1 (32.1–34.5) (24.0–26.2) (24.0–26.2) (29.0–31.4) (23.4–25.6) (29.9–32.3) Diabetes 8.4 (7.8–9.0) 7.9 (7.3–8.5) 5.5 (5.0–6.0) 6 (5.5–6.5) 6.7 (6.1–7.3) 8.6 (8.0–9.2) 6.7 (6.1–7.3) 10.7 (10.0–11.4) 8.6 (6.0–9.2) 7.2 (6.6–7.8) 7.9 (7.3–8.5) Hyperlipidaemia 22.3 21.5 (20.5–22.5) 13.4 16.8 17.1 20.9 17 (16.1–17.9) 25.7 (24.6–26.8) 24.3 (23.2–25.4) 19.3 (18.4–20.2) 20.9 (21.3–23.3) (12.6–14.2) (15.9–17.7) (16.2–18.0) (19.9–21.9) (19.9–21.9) Obesity 7.3 (6.7–7.9) 6.7 (6.1–7.3) 10.4 (9.7–11.1) 7.6 (7.0–8.2) 8.4 (7.8–9.0) 11.4 9.7 (9.0–10.4) 11.4 (10.7–12.1) 6.4 (5.9–6.9) 11.7 (11.0–12.4) 8.1 (7.5–8.7) (10.7–12.1) History of cerebral arterial 10.9 10.9 (10.2–11.6) 3.7 (3.3–4.1) 5.0 (4.5–5.5) 5.8 (5.3–6.3) 6.3 (5.8–6.8) 5.2 (4.7–5.7) 9.3 (8.6–10.0) 10.4 (9.7–11.1) 5.0 (4.5–5.5) 8.6 (8.0–9.3) disease (10.2–11.6) History of ischaemic heart 10.4 (9.7–11.1) 10.6 (9.9–11.3) 5.3 (4.8–5.8) 7.0 (6.4–7.6) 7.6 (7.0–8.2) 8.5 (7.9–9.1) 6.9 (6.3–7.5) 11.5 (10.8–12.2) 11.7 (11.0–12.4) 7.6 (7.0–8.2) 9.6 (8.9–10.2) disease Chronic pulmonary disease 9.5 (8.8–10.2) 9.3 (8.6–10.0) 8.2 (7.6–8.8) 8.5 (7.9–9.1) 8.6 (8.0–9.2) 9.4 (8.7–10.1) 8.5 (7.9–9.1) 11.8 (11.1–12.5) 10.8 (10.1–11.5) 9.0 (8.4–9.6) 9.5 (8–9–10.2) Dementia 2.6 (2.3–2.9) 3.4 (3.0–3.8) 0.6 (0.4–0.8) 0.6 (0.4–0.8) 1.5 (1.2–1.8) 1.1 (0.9–1.3) 0.8 (0.6–1.0) 1.1 (0.9–1.3) 3.4 (3.0–3.8) 0.7 (0.5–0.9) 2.2 (1.9–2.5) GePaRD, Germany (n = 1,715,845) Hypertension 57 (55.9–58.1) 52.3 (51.2–53.4) 45.2 50.3 55.0 66.8 52.7 63.0 (61.8–64.2) 63.0 (61.8–64.2) 55.3 (54.2–56.4) 58.4 (44.2–46.2) (49.2–51.4) (53.9–56.1) (65.6–68.0) (51.6–53.8) (57.3–59.6) Diabetes 18.8 16.3 (15.7–16.9) 13.3 13.9 17.9 26.7 15.8 20.6 (19.9–21.3) 20.5 (19.8–21.2) 16.1 (15.5–16.7) 19.0 (18.2–19.4) (12.8–13.8) (13.3–14.5) (17.3–18.5) (25.9–27.5) (15.2–16.4) (18.4–19.7) Hyperlipidaemia 45.1 41.4 (40.4–42.4) 35.9 39.2 43.7 53.2 41.7 49.9 (48.8–51.0) 50.0 (48.9–51.1) 43.7 (42.7–44.7) 46.3 (44.1–46.1) (35.0–36.8) (38.3–40.1) (42.7–44.7) (52.1–54.3) (40.7–42.7) (45.3–47.3) Obesity 24.9 23.2 (22.5–23.9) 28.3 22.9 26.0 34.1 25.7 27.9 (27.1–28.7) 24.0 (23.3–24.7) 27.2 (26.4–28.0) 25.9 (24.2–25.6) (27.5–29.1) (22.2–23.6) (25.2–26.8) (33.2–35.0) (24.9–26.5) (25.1–26.6) 245 History of cerebral arterial 9.7 (9.2–10.2) 8.6 (8.2–9.0) 4.3 (4.0–4.6) 5.1 (4.8–5.4) 8.3 (7.9–8.7) 9.2 (8.7–9.7) 6.5 (6.1–6.9) 7.5 (7.1–7.9) 10.1 (9.6–10.6) 5.8 (5.4–6.2) 8.5 (8.0–8.9) disease History of ischaemic heart 22.7 19.9 (19.2–20.6) 14.0 17.8 21.4 28.0 18.5 25.2 (24.4–26.0) 26.9 (26.1–27.7) 19.8 (19.1–20.5) 23.2 disease (22.0–23.4) (13.4–14.6) (17.2–18.4) (20.7–22.1) (27.2–28.8) (17.9–19.1) (22.2–23.9) Chronic pulmonary disease 49.8 49.0 (48.0–50.0) 51.7 49.5 50.6 53.5 50.1 52.3 (51.2–53.4) 50.6 (49.5–51.7) 52.5 (51.4–53.6) 50.8 (48.7–50.9) (50.6–52.8) (48.4–50.6) (49.5–51.7) (52.4–54.6) (49.0–51.2) (49.7–51.9) Dementia 5.9 (5.5–6.3) 4.4 (4.1–4.7) 1.6 (1.4–1.8) 2.3 (2.1–2.5) 5.5 (5.1–5.9) 3.9 (3.6–4.2) 3.5 (3.2–3.8) 2.5 (2.3–2.7) 6.3 (5.9–6.7) 2.6 (2.4–2.8) 4.6 (4.3–4.9) EpiChron, Aragon, Spain (n = 211,284) Hypertension 51.2 43.4 (40.6–46.2) 37.7 37.4 51.3 48.6 41.5 39.1 (36.4–41.8) 55.2 (52.0–58.4) 41.7 (39.0–44.4) 44.3 (48.2–54.2) (35.1–40.3) (34.8–40.0) (48.3–54.3) (45.6–51.6) (38.8–44.2) (41.5–47.2) Diabetes 15.2 11.9 (10.4–13.4) 11.3 (9.9–12.7) 9.4 (8.1–10.7) 14.4 14.3 11.7 9.3 (8.0–10.6) 15.9 (14.2–17.6) 11.2 (9.8–12.6) 12.2 (13.5–16.9) (12.8–16.0) (12.7–15.9) (10.2–13.2) (10.7–13.7) Hyperlipidaemia 37.4 34.3 (31.8–36.8) 30.8 31.1 39.1 40.2 34.8 32.7 (30.3–35.1) 39.8 (37.1–42.5) 36.9 (34.3–39.5) 35.3 (34.8–40.0) (28.4–33.2) (28.7–33.5) (36.4–41.8) (37.5–42.9) (32.3–37.3) (32.8–37.9) Obesity 21.2 19.8 (17.9–21.7) 32.1 19.1 21.8 25.3 21.0 21.0 (19.0–23.0) 18.2 (16.4–20.0) 22.5 (20.5–24.5) 21.6 (19.2–23.2) (29.7–34.5) (17.2–21.0) (19.8–23.8) (23.2–27.4) (19.0–23.0) (19.7–23.6) History of cerebral arterial 5.4 (4.4–6.4) 4.3 (3.4–5.2) 2.3 (1.7–2.9) 2.4 (1.7–3.1) 5.2 (4.2–6.2) 4.1 (3.2–5.0) 3.0 (2.3–3.7) 3.3 (2.5–4.1) 5.3 (4.3–6.3) 3.1 (2.3–3.9) 3.9 (3.0–4.7) disease History of ischaemic heart 4.6 (3.7–5.5) 3.7 (2.9–4.5) 2.3 (1.7–2.9) 2.7 (2.0–3.4) 5.2 (4.2–6.2) 3.9 (3.1–4.7) 2.7 (2.0–3.4) 2.5 (1.8–3.2) 5.5 (4.5–6.5) 3.3 (2.5–4.1) 3.6 (2.8–4.4) disease Chronic pulmonary disease 9.4 (8.1–10.7) 8.6 (7.4–9.8) 8.8 (7.5–10.1) 7.8 (6.6–9.0) 9.6 (8.3–10.9) 9.8 (8.5–11.1) 7.9 (6.7–9.1) 8.3 (7.1–9.5) 10.7 (9.3–12.1) 9.2 (7.9–10.5) 9.0 (7.7–10.2) Dementia 2.8 (2.1–3.5) 1.8 (1.2–2.4) 0.8 (0.4–1.2) 0.9 (0.5–1.3) 2.6 (1.9–3.3) 1.5 (1.0–2.0) 1.8 (1.2–2.4) 0.7 (0.3–1.1) 4.0 (3.1–4.9) 1.3 (0.8–1.8) 1.8 (1.2–2.3) SIDIAP, Catalonia, Spain (n = 260,720) Hypertension 33.9 27.9 (25.9–29.9) 23.4 25.5 40.7 33.3 30.1 27.9 (25.9–29.9) 38.0 (35.6–40.4) 25.8 (23.9–27.7) 30.9 (31.7–36.1) (21.5–25.3) (23.6–27.4) (38.3–43.1) (31.1–35.5) (28.0–32.2) (28.8–33.0) Diabetes 12.5 10.2 (9.0–11.4) 9.6 (8.4–10.8) 8.6 (7.5–9.7) 15.2 13.1 11.2 (9.9–12.5) 10.2 (9.0–11.4) 15.2 (13.7–16.7) 9.3 (8.1–10.5) 11.5 (11.1–13.9) (13.7–16.7) (11.7–14.5) (10.2–12.8) Hyperlipidaemia 31.3 26.2 (24.2–28.2) 25.6 26.3 35.5 34.2 31.7 29.5 (27.4–31.6) 34.7 (32.4–37.0) 29.1 (27.0–31.2) 30.1 (29.2–33.4) (23.7–27.5) (24.3–28.3) (33.2–37.8) (32.0–36.4) (29.5–33.9) (28.0–32.2) (continued on next page) Journal of Affective Disorders 249 (2019) 242–252 Table 3 (continued) SSRIs,% (95% CI) SNRIs,% (95% CI) TCA,% (95% CI) NASSA,% (95% Other,% (95% CI) CI) J. Forns, et al. Citalopram Escitalopram Fluoxetine Paroxetine Sertraline Duloxetine Venlafaxine Amitriptyline Mirtazapine Agomelatine Overall Obesity 18.6 14.8 (13.3–16.3) 24.2 15.7 20.2 20.6 19.5 18.7 (17.0–20.4) 16.0 (14.5–17.5) 18.0 (16.4–19.6) 18.4 (16.9–20.3) (22.3–26.1) (14.2–17.2) (18.5–21.9) (18.9–22.3) (17.8–21.2) (16.7–20.0) History of cerebral arterial 5.0 (4.1–5.9) 4.1 (3.3–4.9) 2.2 (1.6–2.8) 2.7 (2.1–3.3) 5.8 (4.9–6.7) 4.1 (3.3–4.9) 3.9 (3.1–4.7) 3.5 (2.8–4.2) 6.4 (5.4–7.4) 3.0 (2.3–3.7) 4.1 (3.4–4.9) disease Chronic pulmonary disease 13.0 11.5 (10.2–12.8) 11.4 11.5 14.1 12.7 12.3 12.3 (11.0–13.6) 15.4 (13.9–16.9) 11.6 (10.3–12.9) 12.6 (11.6–14.4) (10.1–12.7) (10.2–12.8) (12.7–15.5) (11.3–14.1) (11.0–13.6) (11.2–14.0) Dementia 2.6 (2.0–3.2) 2.5 (1.9–3.1) 0.6 (0.3–0.9) 1.3 (0.9–1.7) 2.7 (2.1–3.3) 1.1 (0.7–1.5) 2.0 (1.5–2.5) 0.6 (0.3–0.9) 5.2 (4.3–6.1) 0.9 (0.5–1.3) 2.0 (1.5–2.6) Sweden (n = 1,806,881) Hypertension 45.5 33.1 (32.3–33.9) 25.9 32.1 31.6 40.7 31.8 46.8 (45.8–47.8) 48.0 (47.0–49.0) 36.6 (35.7–37.5) 40.0 (44.5–46.5) (25.2–26.6) (31.3–32.9) (30.8–32.4) (39.8–41.6) (31.0–32.6) (39.1–40.9) Diabetes 9.5 (9.1–9.9) 6.2 (5.8–6.6) 5.0 (4.7–5.3) 5.4 (5.1–5.7) 6.2 (5.8–6.6) 9.9 (9.4–10.4) 6.2 (5.8–6.6) 10.2 (9.7–10.7) 9.8 (9.3–10.3) 6.4 (6.0–6.8) 8.2 (7.8–8.6) Hyperlipidaemia 22.2 14.6 (14.0–15.2) 9.8 (9.3–10.3) 13.5 14.5 19.9 14.1 23.4 (22.7–24.1) 23.3 (22.6–24.0) 14.2 (13.7–14.7) 18.9 (21.5–22.9) (13.0–14.0) (13.9–15.1) (18.4–19.6) (13.6–14.6) (18.3–19.5) Obesity 2.9 (2.7–3.1) 3.5 (3.2–3.8) 5.7 (5.4–6.0) 3.2 (2.9–3.5) 3.4 (3.1–3.7) 5.9 (5.5–6.3) 4.4 (4.1–4.7) 4.7 (4.4–5.0) 3.0 (2.7–3.3) 5.5 (5.2–5.8) 3.7 (3.4–4.0) History of cerebral arterial 8.6 (8.2–9.0) 3.8 (3.5–4.1) 1.7 (1.5–1.9) 2.4 (2.2–2.6) 3.8 (3.5–4.1) 3.5 (3.2–3.8) 2.9 (2.7–3.1) 5.1 (4.8–5.4) 8.5 (8.1–8.9) 2.1 (1.9–2.3) 5.7 (5.3–6.0) disease History of ischaemic heart 10.1 (9.6–10.6) 4.9 (4.6–5.2) 2.3 (2.1–2.5) 4.2 (3.9–4.5) 5.1 (4.8–5.4) 5.7 (5.4–6.0) 4.0 (3.7–4.3) 8.2 (7.8–8.6) 10.8 (10.3–11.3) 3.4 (3.1–3.7) 7.5 (7.1–7.9) disease Chronic pulmonary disease 8.3 (7.9–8.7) 7.4 (7.0–7.8) 7.6 (7.2–8.0) 6.9 (6.5–7.3) 7.3 (6.9–7.7) 9.0 (8.6–9.4) 7.4 (7.0–7.8) 9.3 (8.9–9.7) 9.6 (9.1–10.1) 8.1 (7.7–8.5) 8.3 (7.9–8.7) Dementia 4.2 (3.9–4.5) 1.7 (1.5–1.9) 0.3 (0.2–0.4) 0.6 (0.5–0.7) 1.4 (1.2–1.6) 0.9 (0.8–1.0) 1.0 (0.9–1.1) 0.7 (0.6–0.8) 5.2 (4.9–5.5) 0.4 (0.3–0.5) 2.6 (2.3–2.8) CI = confidence interval, NASSA = noradrenaline and specific serotonergic antidepressant, SNRI = serotonin-norepinephrine reuptake inhibitor, SSRI = selective serotonin reuptake inhibitor, TCA = tricyclic anti- depressant. 246 Journal of Affective Disorders 249 (2019) 242–252 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 Except in Denmark where the cumulative incidence was lower (10 per hospitalised in the previous year, followed by initiators of mirtazapine 1000 population), amitriptyline had a similar cumulative incidence in (47% of patients with at least one hospitalisation). all populations (ranging from 21 to 29 per 1000 population). The cu- mulative incidence for mirtazapine was the highest in Denmark, Ger- 3.5. Patterns of antidepressant use among initiators many (where it was the second most commonly used antidepressant), and Sweden (where it was the third). Finally, the use of agomelatine The recent use of any antidepressant during the 12 months prior to was low in all populations, with a cumulative incidence below 10 per the current treatment episode is presented in Fig. 1. In all populations, 1000 population in all data sources. citalopram initiators were among the patients with the highest pro- portions of no antidepressant use in the 12 months prior to the current 3.3. Baseline comorbidities of antidepressant initiators treatment episode. Amitriptyline had the highest proportion with no use in Aragon and Catalonia (Spain) and in Germany. Conversely, the The most prevalent comorbidities at baseline (percentage and 95% two SNRIs (duloxetine and venlafaxine) and agomelatine were the an- confidence interval) in the study antidepressant initiators were hy- tidepressants with the higher proportion of patients who used two or pertension, diabetes, hyperlipidaemia, and obesity (Table 3). The more antidepressants during the 12 months prior to the current treat- highest prevalence of hypertension was found among initiators of ment episode (Fig. 1). Among the five populations, the overall pro- mirtazapine in Aragon (55%) and Catalonia (38%) (Spain) and Sweden portion of initiators with no previous use of other study antidepressants (48%) and among initiators of amitriptyline in Denmark (38%). In in the 12 months prior to the current treatment episode was highest in Germany, the highest prevalence of hypertension was found equal Spain, Germany, and Denmark and lowest in Sweden. among initiators of amitriptyline and mirtazapine (63%). A similar Patterns regarding duration of the first treatment episode of current pattern, although with lower prevalence compared with hypertension, use and percentage of all treatment episodes of current use with com- was observed for other comorbidities. Conversely, the percentage of bined use of other antidepressants during the study period are pre- patients with obesity at baseline was higher among fluoxetine initiators sented in Table 5. For all study antidepressants, the duration of the first and lower for mirtazapine initiators. treatment episode of current use was slightly longer in Catalonia (Spain) (range, 4 months for amitriptyline to 10 months for venla- 3.4. Hospitalisations in the 12 months prior to start date among faxine) and shorter in Germany (range, 2 months for amitriptyline to 4 antidepressant initiators months for fluoxetine) than in the other populations. The median duration of the first index episode was longer for venlafaxine initiators The number of hospitalisations among initiators of study anti- and shorter for amitriptyline initiators in all populations except in depressants during the 12 months before the study start varied among Sweden. The highest percentage of patients in the five populations who populations, although a common pattern emerged. Mirtazapine in- were using combined antidepressants during their current treatment itiators were more frequently hospitalised compared with initiators of episode was found in initiators of agomelatine, followed by venlafaxine the remaining study antidepressants in Aragon and Catalonia, Spain, and duloxetine initiators (in Denmark, amitriptyline also had a high and Sweden (percentages of at least one hospitalisation were 23%, percentage of multiple use). 17%, and 33%, respectively) (Table 4). Although initiators of ami- triptyline had the highest proportion of patients who had been hospi- 4. Discussion talised in the previous year in Denmark (37% of patients had at least 1 hospitalisation), the proportion of patients with recent hospitalisations In this study, we characterised more than 4.8 million initiators of 10 was also high for initiators of mirtazapine (33% of patients with at least antidepressants in Spain, Germany, Denmark, and Sweden. Overall, 1 hospitalisation). Similarly, in Germany initiators of duloxetine and citalopram was the most prescribed study antidepressant, followed by escitalopram (56% and 48% of patients with at least one hospitalisa- mirtazapine. Agomelatine (one of the most recent study antidepressants tion, respectively) had the highest proportion of patients who had been available on the market) and paroxetine, were the least commonly Table 4 Number of hospitalisations among study antidepressant initiators of each of the study antidepressants during the 12 months before the cohort inclusion date, by data source. SSRIs SNRIs TCA NASSA Other Citalopram Escitalopram Fluoxetine Paroxetine Sertraline Duloxetine Venlafaxine Amitriptyline Mirtazapine Agomelatine Denmark 0 70.2% 68.4% 78.6% 76.9% 76.2% 73.6% 76.5% 62.9% 67.2% 77.8% 1 17.6% 17.7% 14.3% 14.4% 15.2% 15.9% 14.8% 20.3% 18.0% 14.2% 2+ 12.1% 13.8% 7.1% 8.7% 8.6% 10.5% 8.7% 16.8% 14.8% 8.0% GePaRD 0 60.4% 52.0% 68.1% 65.2% 56.9% 44.1% 55.3% 56.7% 52.9% 53.9% 1+ 39.6% 48.0% 31.9% 34.8% 43.1% 55.9% 44.7% 43.3% 47.1% 46.1% EpiChron, Aragon, Spain 0 84.2% 85.7% 89.5% 88.9% 82.7% 84.9% 86.1% 89.8% 76.7% 86.5% 1 10.8% 10.1% 7.7% 8.3% 11.5% 10.6% 10.1% 7.9% 15.0% 9.8% 2+ 5.0% 4.2% 2.8% 2.9% 5.8% 4.5% 3.9% 2.4% 8.3% 3.7% SIDIAP, Catalonia, Spain 0 88.4% 88.8% 90.9% 91.4% 87.3% 88.7% 88.7% 90.1% 83.1% 91.4% 1 8.7% 8.2% 7.3% 6.8% 9.4% 8.7% 8.8% 7.8% 11.7% 6.5% 2+ 3.0% 3.0% 1.9% 1.8% 3.3% 2.7% 2.5% 2.1% 5.2% 2.1% Sweden 0 73.8% 75.9% 80.2% 80.0% 79.3% 71.8% 76.7% 75.2% 66.8% 73.3% 1 13.8% 12.8% 11.8% 11.6% 12.2% 14.8% 13% 13.4% 15.8% 13.6% 2+ 12.4% 11.2% 8.0% 8.4% 8.5% 13.4% 10.3% 11.4% 17.5% 13.1% NA = not applicable, NASSA = noradrenaline and specific serotonergic antidepressant, SNRI = serotonin-norepinephrine reuptake inhibitor, SSRI = selective ser- otonin reuptake inhibitor, TCA = tricyclic antidepressant. 247 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 Fig. 1. Proportion of study antidepressants initiators by antidepressant treatment in the 12 months before drug initiation (no previous treatment, one antidepressant, two or more antidepressants) NASSA = noradrenaline and specific serotonergic antidepressant, SNRI = serotonin-norepinephrine reuptake inhibitor, SSRI = selective serotonin reuptake inhibitor, TCA = tricyclic antidepressant. prescribed antidepressants. Mirtazapine initiators were older compared response is not adequate, the guidelines suggest a different SSRI agent to initiators of other antidepressants. Consequently, mirtazapine in- or a newer generation antidepressant like mirtazapine, which was the itiators had the highest percentage of comorbidities at baseline and also second most prescribed antidepressant in Denmark, Germany, and more hospitalisations in the 12 months prior to start date. Fluoxetine Sweden. It was however unexpected that mirtazapine initiators in all was the most common study antidepressant among younger and heal- populations but Sweden had a low percentage of prior use of other thier initiators. Finally, citalopram and amitriptyline were the anti- antidepressants in the 12 months prior to cohort inclusion, suggesting depressants with the lowest percentages of initiators who had used that mirtazapine may to some extent be used as a first-line agent in the other antidepressant drugs in the 12 months prior to the current studied populations. The faster onset of action and the more sedative treatment episode, while agomelatine and SNRIs (duloxetine and ven- effect of mirtazapine compared with SSRIs in the acute phase of de- lafaxine) initiators had the highest percentage of antidepressants use in pression might explain these results (Watanabe et al., 2011). Re- the year prior to the current treatment episode. imbursement policies vary by antidepressant in each of the partici- The results of the present study, which highlight citalopram as the pating countries, and this could potentially explain the observed most commonly prescribed among study antidepressant initiators in all variability in patterns of use of antidepressants. As an example, the populations but in Aragon (Spain), where escitalopram was the most study revealed differences in prescriptions in the two Spanish popula- often prescribed antidepressant, are in accordance with clinical guide- tions, with escitalopram being more frequently prescribed in Aragon, lines to manage major depressive disorder in adults (Cipriani et al., while citalopram was the number one option in Catalonia. 2018; National Institute for Health and Clinical Excellence, 2018b). The Overall, more than 50% of the antidepressant initiators receiving clinical guidelines for depression suggest using an SSRI (mostly citalo- SSRIs (mostly citalopram and sertraline) had no prior use of other an- pram, fluoxetine, and sertraline) as first-line treatment in adults tidepressants, and a very low proportion of these patients used two or (National Institute for Health and Clinical Excellence, 2018b). If clinical more antidepressants in the 12 months prior to current treatment 248 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 Fig. 1. (continued) Fig. 1. (continued) 249 J. Forns, et al. Table 5 Duration of treatment and multiple use of study antidepressants, by study population. SSRIs SNRIs TCA NASSA Others Citalopram Escitalopram Fluoxetine Paroxetine Sertraline Duloxetine Venlafaxine Amitriptyline Mirtazapine Agomelatine Denmark Duration (months) of first treatment episode in current use, 7.0 (4.6–15.4) 5.6 (2.5–12.9) 6.5 (5.1–13.5) 5.8 (3.9–13.0) 6.7 6.5 (2.5–16.3) 8.2 (5.2–18.7) 3.5 (3.5–5.9) 4.8 (2.6–9.4) 5.6 (2.5–13.1) median (IQR) (5.4–13.9) Percentage of all treatment episodes (current use) with 0 (0–30.0) 9.5 (0–62.6) 9.2 (0–60.7) 0 (0–46.7) 0 (0–42.5) 34.8 (1.2–99.4) 22.0 (0–77.6) 40.8 (0–100.0) 16.3 (0–100.0) 69.2 (7.7–100.0) combined use of other antidepressant, median (IQR) GePaRD Duration (months) of first treatment episode in current use, 3.0 (2.0–6.1) 3.0 (2.0–7.1) 3.9 (2.1–7.2) 3.0 (2.0–6.2) 3.5 (2.0–7.9) 2.3 (1.8–5.0) 3.2 (1.9–7.4) 1.8 (1.5–2.4) 2.1 (1.6–3.1) 2.7 (2.2–5.6) median (IQR) Percentage of all treatment episodes (current use) with 0.0 (0.0–51.0) 18.6 (0.0–81.9) 1.7 (0.0–62.8) 2.8 (0.0–70.5) 11.0 9.3 (0.0–86.7) 26.4 (0.0–94.5) 0.0 (0.0–0.0) 0.0 (0.0–82.0) 44.0 (0.0–100.0) combined use of other antidepressant, median (IQR) (0.0–78.2) EpiChron, Aragon, Spain Duration (Months) of first treatment episode in current use, 3.3 (2.2–7.9) 3.4 (2.2–7.9) 3.5 (2.2–7.7) 3.2 (2.2–7.3) 3.5 (2.3–8.3) 3.6 (2.2–9.8) 4.0 (2.3–10.3) 3.3 (2.1–5.1) 3.3 (2.3–7.3) 3.3 (2.2–7.7) 250 median (IQR) Percentage of all treatment episodes (current use) with 0 (0–14.2) 0 (0–6.9) 0 (0–10.7) 0 (0–10.7) 0 (0–22.5) 0 (0–33.5) 5.1 (0–50.8) 0 (0–1.4) 5.3 (0–67.8) 25.0 (0–74.5) combined use of other antidepressant, median (IQR) SIDIAP, Catalonia, Spain Duration (months) of first treatment episode in current use, 7.6 (3.2–17.5) 8.4 (3.2–19.6) 7.3 (3.3–16.9) 7.3 (3.0–15.9) 7.3 8.2 (3.2–19.8) 9.7 (3.5–22.2) 3.7 (2.1–9.5) 6.3 (2.4–15.0) 8.0 (3.3–18.8) median (IQR) (3.0–16.3) Percentage of all treatment episodes (current use) with 0.0 (0.0–10.0) 0.0 (0.0–19.9) 0.0 (0.0–15.6) 0.0 (0.0–10.6) 0.0 3.9 (0.0–49.3) 9.3 (0.0–63.1) 0.0 (0.0–0.0) 8.3 (0.0–100.0) 18.8 (0.0–100.0) combined use of other antidepressant, median (IQR) (0.0–19.4) Sweden Duration (months) of first treatment episode in current use, 4.6 (2.6–10.3) 3.3 (2.1–7.2) 4.5 (2.3–7.7) 4.6 (2.1–7.9) 4.6 (2.3–8.8) 2.2 (1.7–4.5) 2.8 (1.6–4.7) 4.6 (2.6–5.9) 3.2 (2.1–5.7) 2.2 (1.4–4.1) median (IQR) Percentage of all treatment episodes (current use) with 0 (0–14.5) 0 (0–50.2) 0 (0–42.6) 0 (0–39.1) 0 (0–25.0) 17.4 (0–77.2) 19.2 (0–72.9) 0 (0–14.8) 15 (0–95.9) 69.2 (0–100.0) combined use of other antidepressant, median (IQR) CI = confidence interval, IQR = interquartile range, NA = not applicable, NASSA = noradrenaline and specific serotonergic antidepressant, SNRI = serotonin-norepinephrine reuptake inhibitor, SSRI = selective ser- otonin reuptake inhibitor, TCA = tricyclic antidepressant. Journal of Affective Disorders 249 (2019) 242–252 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 episode, except in Sweden. Also, most SSRI initiators in the five po- of its limitations. First, although the data sources provide detailed in- pulations did not have combined use with other antidepressants during formation on filled prescriptions, which can be considered complete, the current treatment episode. Both results suggest that, in general, this information may not reflect actual use. However, information on citalopram and to a minor degree sertraline, escitalopram, paroxetine, dispensing more closely reflects actual use by the patients than in- and fluoxetine were mostly used as first-line treatment and used in formation obtained by prescribing data (Pottegard et al., 2014). Second, monotherapy. On the other hand, the two SNRIs included in the study, the present study used a new-user design and then focused on initiators duloxetine and venlafaxine, as well as agomelatine were seemingly of each of the study antidepressants only. Therefore, prevalent users mostly used as second-line treatment, in agreement with clinical have not been evaluated. Third, the lack of information on the indica- guidelines for major depression (National Institute for Health and tion of medications in the present study hampered the possibility of Clinical Excellence, 2018b). This interpretation is also supported by the studying the differences in the indications across the different study high proportion of patients who used two or more antidepressants in antidepressants and data sources. For this reason, discussion on the the 12 months prior to start date as well as by the high percentage of present study focuses on depression and not on other conditions that patients who used other antidepressants concomitantly during the can be treated with the study antidepressants. Fourth, only a selected study period. However, of the antidepressants included in this study, group of antidepressants have been included in the present study, based duloxetine, venlafaxine, and agomelatine are the ones introduced in the on the list of antidepressants included in the Agomelatine PASS European Union market more recently, which could explain these re- (Pladevall, 2018), and therefore the results of this study did not include sults. all available antidepressants in the ATC N06A group. Finally, the het- The main results discussed in the present study are focused on erogeneity of health care systems and of the type of data across data management of depression, but all the study antidepressants are also sources needs to be considered when evaluating comorbidities asso- used in other conditions, which may explain some of the results ob- ciated with the different antidepressants. Information recorded in SI- served in the present study. For example, the TCA amitriptyline was DIAP (Catalonia, Spain) and EpiChron (Aragon, Spain) data sources is mainly used as monotherapy, and use of other antidepressants in the 12 based on primary care and hospital discharge electronic medical re- months prior to the current treatment episode was uncommon. cords, information recorded in the GePaRD (Germany) data source is Amitriptyline is recommended as first-line treatment in the manage- based on insurance claims from ambulatory care visits and hospitali- ment of neuropathic pain (Fornasari, 2017; National Institute for Health sations, and information in Denmark and Sweden registers is based on and Clinical Excellence, 2018a) and can also be used in the context of hospital diagnoses only. migraine, stress, and sleep disorders (Aarts et al., 2016; Moore et al., On the other hand, the present study has a number of strengths. The 2012, 2015). For these indications, amitriptyline can be used as present study presented data from 2009 to 2014, the most recent period monotherapy, which accords with the observations from Germany, reported in the literature. In addition, it is one of the largest drug uti- Sweden, and Spain, or in combination with pregabalin/gabapentin, lization studies on the use of antidepressants in Europe, including al- opioids, or SNRIs (Holbech et al., 2017). Finally, the two SNRIs in the most 5 million initiators in Spain, Germany, Denmark, and Sweden study, duloxetine and venlafaxine, as well as the SSRIs are approved covering a base population of more than 36 million people. Also, we treatments for anxiety disorders (Aarts et al., 2016). have studied a number of important comorbidities at baseline. Finally, The present study revealed that initiators of study antidepressants we have well-defined populations with prospectively collected data. frequently had several comorbidities, hypertension being the most frequent. Mirtazapine was the most prescribed antidepressant among 6. Conclusions patients with a higher percentage of comorbidities at baseline, which is probably related to the fact that mirtazapine initiators were the oldest Results of this study conducted in Denmark, Germany, Spain, and initiators in all populations (median age range, 54–67 years). The un- Sweden including 4.8 million initiators of study antidepressants suggest ique adverse event profile of mirtazapine including sedation and in- that citalopram and mirtazapine are the most commonly used anti- crease in appetite may justify its use in older patients compared with depressants, followed by other SSRIs (sertraline, escitalopram, parox- other antidepressants (Watanabe et al., 2011). It is important to note etine, and fluoxetine). Mirtazapine was the second most used anti- that mirtazapine and amitriptyline, particularly in the Nordic countries, depressant in most of the populations. Agomelatine, which is commonly were the most commonly prescribed antidepressants among patients used as second-line treatment, and paroxetine were the least used an- presenting comorbidities such as hypertension, diabetes, or hyperlipi- tidepressants in the participating populations. Mirtazapine was the daemia, which does not reflect the clinical guidelines that recommend antidepressant most commonly prescribed among older study anti- the use of citalopram or sertraline in these patient groups depressant initiators and was the one with the highest percentage of (National Institute for Health and Clinical Excellence, 2018b) which comorbidities in patients at baseline and with the highest number of may be partly explained for the approved use in other conditions. hospitalisations. Fluoxetine was commonly prescribed among young Conversely, fluoxetine initiators were younger compared to mirtaza- patients with a low percentage of comorbidities at baseline. pine initiators and had the lowest percentage of comorbidities at baseline among all study antidepressants. These results are likely as- Acknowledgements sociated with the high efficacy of fluoxetine in reducing depressive symptoms in children and adolescents (Cipriani et al., 2016). For obe- The authors would like to thank Estel Plana, MSc (Epidemiology, sity, the percentage was higher among fluoxetine initiators and lower RTI Health Solutions, Barcelona, Spain); Carla Franzoni, BSc among mirtazapine initiators, which is in accordance with previous (Epidemiology, RTI Health Solutions, Barcelona, Spain); Niklas evidence that relates mirtazapine to a greater risk of weight gain and Schmedt, PhD (InGef - Institute for Applied Health Research, Berlin, fluoxetine with weight loss (Gafoor et al., 2018). Because we were not Germany), Marieke Niemeyer, MSc (Leibniz Institute for Prevention able to differentiate whether differences in prevalence of comorbidities Research and Epidemiology - BIPS, Bremen, Germany), Sandra Ulrich, between countries reflected an actual variation or differences in the MSc (Leibniz Institute for Prevention Research and Epidemiology - type of data available in each population, this could not be further BIPS, Bremen, Germany), Morten Olesen (Clinical Pharmacology and interpreted. Pharmacy, Department of Public Health, University of Southern Denmark, Odense, Denmark), and Martin Thomsen Ernst (Clinical 5. Limitations Pharmacology and Pharmacy, Department of Public Health, University of Southern Denmark, Odense, Denmark; and OPEN, Odense Patient The results of the present study should be interpreted in the context data Explorative Network, Odense University Hospital, Odense 251 J. Forns, et al. Journal of Affective Disorders 249 (2019) 242–252 Denmark). The authors also would like to thank the German statutory the online version, at doi:10.1016/j.jad.2019.02.010. health insurance providers that provided data for the study in GePaRD, namely the AOK Bremen/Bremerhaven, the DAK-Gesundheit, and Die References Techniker. Aarts, N., Noordam, R., Hofman, A., Tiemeier, H., Stricker, B.H., Visser, L.E., 2016. Self-re- ported indications for antidepressant use in a population-based cohort of middle-aged and Disclosure statement elderly. Int. J. Clin. Pharm. 38, 1311–1317. Abbing-Karahagopian, V., Huerta, C., Souverein, P.C., de Abajo, F., Leufkens, H.G., Slattery, J., Joan Forns, Miguel Cainzos-Achirica, Susana Perez-Gutthann, and Alvarez, Y., Miret, M., Gil, M., Oliva, B., Hesse, U., Requena, G., de Vries, F., Rottenkolber, M., Schmiedl, S., Reynolds, R., Schlienger, R.G., de Groot, M.C., Klungel, O.H., van Staa, Manel Pladevall are employees of RTI Health Solutions, a unit of RTI T.P., van Dijk, L., Egberts, A.C., Gardarsdottir, H., De Bruin, M.L., 2014. Antidepressant International, a non-profit organisation that conducts work for gov- prescribing in five European countries: application of common definitions to assess the ernment, public, and private organisations, including pharmaceutical prevalence, clinical observations, and methodological implications. Eur. J. Clin. Pharmacol. 70, 849–857. companies. Bauer, M., Monz, B.U., Montejo, A.L., Quail, D., Dantchev, N., Demyttenaere, K., Garcia- Johan Reutfors, David Hägg, and Lena Brandt are employees of the Cebrian, A., Grassi, L., Perahia, D.G., Reed, C., Tylee, A., 2008. Prescribing patterns of antidepressants in Europe: results from the Factors Influencing Depression Endpoints Centre for Pharmacopidemiology, which receives grants from several Research (FINDER) study. Eur. Psychiatry 23, 66–73. entities (pharmaceutical companies, regulatory authorities, contract Cipriani, A., Furukawa, T.A., Salanti, G., Chaimani, A., Atkinson, L.Z., Ogawa, Y., Leucht, S., research organisations) for the performance of drug safety and drug Ruhe, H.G., Turner, E.H., Higgins, J.P.T., Egger, M., Takeshima, N., Hayasaka, Y., Imai, H., Shinohara, K., Tajika, A., Ioannidis, J.P.A., Geddes, J.R., 2018. Comparative efficacy and utilisation studies. acceptability of 21 antidepressant drugs for the acute treatment of adults with major de- Anton Pottegård reports participation in research projects funded by pressive disorder: a systematic review and network meta-analysis. Lancet 391, 1357–1366. Cipriani, A., Zhou, X., Del Giovane, C., Hetrick, S.E., Qin, B., Whittington, C., Coghill, D., Zhang, Alcon, Almirall, Astellas, Astra-Zeneca, Novo Nordisk, LEO Pharma, Y., Hazell, P., Leucht, S., Cuijpers, P., Pu, J., Cohen, D., Ravindran, A.V., Liu, Y., Michael, and Servier, all with funds paid to the institution where he was em- K.D., Yang, L., Liu, L., Xie, P., 2016. Comparative efficacy and tolerability of anti- ployed (no personal fees) and with no relation to the work reported in depressants for major depressive disorder in children and adolescents: a network meta- analysis. Lancet 388, 881–890. this paper. Fornasari, D., 2017. Pharmacotherapy for neuropathic pain: a review. Pain Ther. 6, 25–33. Jesper Hallas has participated in research projects funded by Gafoor, R., Booth, H.P., Gulliford, M.C., 2018. Antidepressant utilisation and incidence of Novartis, Pfizer, Menarini, MSD, Nycomed, Leo Pharmaceuticals, weight gain during 10 years' follow-up: population based cohort study. BMJ 361, k1951. Holbech, J.V., Jung, A., Jonsson, T., Wanning, M., Bredahl, C., Bach, F.W., 2017. Combination Almirall, Servier, Astellas, and Alkabello with grants paid to the in- treatment of neuropathic pain: Danish expert recommendations based on a Delphi process. stitution where he was employed. He has personally received fees for J. Pain Res. 10, 1467–1475. Jobski, K., Kollhorst, B., Garbe, E., Schink, T., 2017a. The risk of ischemic cardio- and cere- teaching or consulting from the Danish Association of Pharmaceutical brovascular events associated with oxycodone-naloxone and other extended-release high- Manufacturers and from Pfizer and Menarini. potency opioids: a nested case-control study. Drug Saf. 40, 505–515. Maja Hellfritzsch has received speaker honorarium fees from Jobski, K., Schmedt, N., Kollhorst, B., Krappweis, J., Schink, T., Garbe, E., 2017b. Characteristics and drug use patterns of older antidepressant initiators in Germany. Eur. J. Bristol-Myers Squibb and Pfizer and a travel grant from LEO Pharma. Clin. Pharmacol. 73, 105–113. Tania Schink, Tammo Reinders, and Bianca Kollhorst, as employees Ludvigsson, J.F., Andersson, E., Ekbom, A., Feychting, M., Kim, J.L., Reuterwall, C., Heurgren, of the Leibniz Institute for Prevention Research and Epidemiology – M., Olausson, P.O., 2011. External review and validation of the Swedish national inpatient register. BMC Public Health 11, 450. BIPS, worked on projects funded by pharmaceutical companies un- Moore, R.A., Derry, S., Aldington, D., Cole, P., Wiffen, P.J., 2012. Amitriptyline for neuropathic related to this study. pain and fibromyalgia in adults. Cochrane Database Syst. Rev. 12, Cd008242. Moore, R.A., Derry, S., Aldington, D., Cole, P., Wiffen, P.J., 2015. Amitriptyline for neuropathic Alexandra Prados-Torres and Beatriz Poblador-Plou are members of pain in adults. Cochrane Database Syst. Rev., Cd008242. the EpiChron Research Group on Chronic Diseases of the Aragon Health National Institute for Health and Clinical Excellence, N., 2018a. Neuropathic pain in adults: Sciences Institute (IACS), ascribed to IIS Aragón, and do not have any pharmacological management in non-specialist settings. National Institute for Health and Clinical Excellence, N., 2018b. NICE clinical guideline. conflict of interest with this project. Depression in adults: recognition and management. Rosa Morros and Maria Giner-Soriano, as employees of IDIAPJGol, Noordam, R., Aarts, N., Verhamme, K.M., Sturkenboom, M.C., Stricker, B.H., Visser, L.E., 2015. Prescription and indication trends of antidepressant drugs in the Netherlands between and Jordi Cortés, as ex-employee of IDIAPJGol, worked on other pro- 1996 and 2012: a dynamic population-based study. Eur. J. Clin. Pharmacol. 71, 369–375. jects funded by pharmaceutical companies in their institution that were Pigeot, I., Ahrens, W., 2008. Establishment of a pharmacoepidemiological database in Germany: not related to this study and without personal profit. methodological potential, scientific value and practical limitations. Pharmacoepidemiol. Drug Saf. 17, 215–223. Emmanuelle Jacquot and Nicolas Deltour are employees of Les Pladevall, M., 2018. Post-authorisation safety study of agomelatine and the risk of hospitali- Laboratoires Servier. sation for acute liver injury. Poluzzi, E., Piccinni, C., Sangiorgi, E., Clo, M., Tarricone, I., Menchetti, M., De Ponti, F., 2013. Trend in SSRI-SNRI antidepressants prescription over a 6-year period and predictors of Funding poor adherence. Eur. J. Clin. Pharmacol. 69, 2095–2101. Pottegard, A., Christensen, R., Houji, A., Christiansen, C.B., Paulsen, M.S., Thomsen, J.L., Hallas, J., 2014. Primary non-adherence in general practice: a Danish register study. Eur. J. The agomelatine post-authorisation study was funded by Les Clin. Pharmacol. 70, 757–763. Laboratoires Servier under a contract granting independent publication Pottegard, A., Schmidt, S.A., Wallach-Kildemoes, H., Sorensen, H.T., Hallas, J., Schmidt, M., rights to the research team. The present manuscript was funded by 2016. Data resource profile: the Danish National Prescription Registry. Int. J. Epidemiol. Prados-Torres, A., Poblador-Plou, B., Gimeno-Miguel, A., Calderon-Larranaga, A., Poncel-Falco, internal resources at each institution. A., Gimeno-Feliu, L.A., Gonzalez-Rubio, F., Laguna-Berna, C., Marta-Moreno, J., Clerencia- Sierra, M., Aza-Pascual-Salcedo, M., Bandres-Liso, A.C., Coscollar-Santaliestra, C., Pico- Author contributions Soler, V., Abad-Diez, J.M., 2018. Cohort profile: the epidemiology of chronic diseases and multimorbidity. EpiChron Cohort Study Int. J. Epidemiol. Schmidt, M., Schmidt, S.A., Sandegaard, J.L., Ehrenstein, V., Pedersen, L., Sorensen, H.T., 2015. Authors J.F., M.P., E.J., N.D., and S.P. planned the study. A.P., T.R., The Danish National Patient Registry: a review of content, data quality, and research po- tential. Clin. Epidemiol. 7, 449–490. B.P., R.M., L.B., M.C., M.H., T.S., A.P.T., M.G., D.H., J.H., and J.C. SIDIAP, 2014. Database. General details. undertook the statistical analysis of the different data sources. J.F., Watanabe, N., Omori, I.M., Nakagawa, A., Cipriani, A., Barbui, C., Churchill, R., Furukawa, M.P., S.P. and J.R. completed the interpretation of the results with T.A., 2011. Mirtazapine versus other antidepressive agents for depression. Cochrane Database Syst. Rev., Cd006528. contributions from all authors. J.F., M.C., M.P., S.P. and J.R. wrote the Wettermark, B., Hammar, N., Fored, C.M., Leimanis, A., Otterblad Olausson, P., Bergman, U., first draft of the manuscript. All authors contributed to and have ap- Persson, I., Sundström, A., Westerholm, B., Rosén, M., 2007. The new Swedish Prescribed proved the final manuscript. Drug Register–opportunities for pharmacoepidemiological research and experience from the first six months. Pharmacoepidemiol. Drug Saf. 16, 726–735. Zimmerman, M., Posternak, M., Friedman, M., Attiullah, N., Baymiller, S., Boland, R., Supplementary materials Berlowitz, S., Rahman, S., Uy, K., Singer, S., 2004. Which factors influence psychiatrists' selection of antidepressants? Am. J. Psychiatry 161, 1285–1289. Supplementary material associated with this article can be found, in 252