- Email: [email protected]
Use of antiepileptic or benzodiazepine medication and suicidal ideation — The Northern Finland Birth Cohort 1966
Epilepsy & Behavior 46 (2015) 198–204
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Use of antiepileptic or benzodiazepine medication and suicidal ideation — The Northern Finland Birth Cohort 1966 I. Rissanen a,b,⁎, E. Jääskeläinen a,b,c, M. Isohanni a,d, H. Koponen e,f, H. Ansakorpi b,g, J. Miettunen a,b,c,d a
Department of Psychiatry, Institute of Clinical Medicine, P.O. Box 5000, 90014, University of Oulu, Oulu, Finland Medical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland University of Oulu, Institute of Health Sciences, P.O. Box 5000, FIN-90014, University of Oulu, Oulu, Finland d Department of Psychiatry, Oulu University Hospital, P.O. Box 26, FIN-90029, OYS, Oulu, Finland e Department of Psychiatry, Institute of Clinical Medicine, University of Helsinki, Helsinki, Finland f Department of Psychiatry, Helsinki University Central Hospital, Helsinki, Finland g Department of Neurology, Oulu University Hospital, University of Oulu, Oulu, Finland b c
a r t i c l e
i n f o
Article history: Received 31 October 2014 Revised 20 January 2015 Accepted 1 March 2015 Available online 29 April 2015 Keywords: Antiepileptic drugs Benzodiazepines Suicidal ideation Cohort study Depression Anxiety
a b s t r a c t Both antiepileptic drugs (AEDs) and benzodiazepines (BZDs) have previously been associated with an increased risk of suicidality. Our aim was to study the association between the use of conventional AEDs and BZDs and suicidal ideation in a large population-based cohort. Information on the medications used in the Northern Finland Birth Cohort 1966 was collected from the subjects at the age of 31 years, using a postal questionnaire (N = 8211). The presence of suicidal ideation and other symptoms of depression and anxiety was assessed via the Hopkins Symptom Checklist — 25 questionnaire. The associations between medications and suicidal ideation were studied in different diagnostic groups and adjusted for symptoms of depression and anxiety. No difference was observed in suicidal ideation between AED users (n = 54) and nonusers (n = 8157). Subjects using BZDs (n = 147) had greater suicidal ideation compared with nonusers (n = 8064). Antiepileptic drug and benzodiazepine users more often exhibited other depression and anxiety symptoms. After adjustment for these symptoms, both AED and BZD users had less suicidal ideation compared with nonusers. In conclusion, in this populationbased cohort, neither the use of AEDs nor that of BZDs was found to be associated with increased suicidal ideation when the symptoms of depression and anxiety were taken into account. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Suicidality is a severe medical condition related not only to several psychiatric disorders, including depression and anxiety [1], but also to epilepsy [2–4]. Suicidality can be divided into suicidal behavior (attempted or committed suicide) and suicidal ideation. Suicidal ideation is usually considered to be a risk factor for suicidal behavior [5–7]. A previous study found the lifetime prevalence of suicidal ideation in people with epilepsy to be 25% [8]. Patients with epilepsy also have a high comorbidity of psychiatric disorders [9]. A high rate of suicidal ideation and behavior among patients with epilepsy associates with comorbid psychiatric disorders, particularly depression and anxiety disorders [3,4]. In 2008, the United States Food and Drug Administration (FDA) mandated a black box warning on antiepileptic drugs (AEDs) because
⁎ Corresponding author at: Institute of Clinical Medicine, Department of Psychiatry, P.O. BOX 5000, 90014, University of Oulu, Oulu, Finland. E-mail address: [email protected].fi (I. Rissanen).
http://dx.doi.org/10.1016/j.yebeh.2015.03.001 1525-5050/© 2015 Elsevier Inc. All rights reserved.
of the increased risk of suicidality [10]. This warning was based on a meta-analysis that revealed a 1.8-fold risk of suicidal ideation or suicidal behavior in AEDs. The FDA study has been widely discussed, and the reliability of the results has been questioned. In contrast to the FDA study, a study by Arana et al. [11] suggested that AEDs were not associated with an increased risk of suicidality in people with epilepsy but were associated with an increased risk of suicidality in people with depression or in people without epilepsy, depression, or bipolar disorder. Benzodiazepine drugs (BZDs) not only are generally used as anxiolytics or hypnotics and sedatives but also have anticonvulsant properties, and some can be used as add-on therapy for epilepsy. Regular use of hypnotics has been associated with increased all-cause mortality and suicide mortality [12]. The use of BZDs has previously been associated with an increased risk of suicide attempts, especially among individuals who had not recently received antidepressant medication [13]. In addition, the use of sedative–hypnotic drugs has been associated with an increased risk of suicidal ideation, plans, and attempts, even when adjusted for demographic characteristics, physical and mental disorders, and sleep disturbances [14]. By contrast, in a meta-analysis studying the use of alprazolam (a benzodiazepine derivate) in the
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
treatment of depression, alprazolam bore a similar risk of the emergence and worsening of suicidal ideation as a placebo, but it showed a higher rate of improvement compared with the placebo in the case of suicidal ideation [15]. Previous studies of AEDs or BZDs and suicidal ideation have been unable to control their findings for other psychiatric symptoms and diagnoses. Few population-based studies have been performed on this subject, although such studies are optimal for estimating the adverse effects of medications [16]. Our aim was to study the associations between AEDs or BZDs and suicidal ideation among groups of different diagnoses or symptoms in a large population-based cohort. Our hypothesis was that both AEDs and BZDs relate not only to increased suicidal ideation but also to other symptoms of depression and anxiety.
2. Methods 2.1. The Northern Finland Birth Cohort 1966 The Northern Finland Birth Cohort 1966 (NFBC 1966) is a populationbased sample containing data on 12,058 babies born alive in the Finnish provinces of Oulu and Lapland with an expected date of birth in 1966 [17]. Since pregnancy, prospective data have been collected on, e.g., the cohort members' health, socioeconomic conditions, families, and living. Permission to gather data was obtained from the Ministry of Social and Health Affairs, and the study was approved by the Ethical Committee of the Northern Ostrobothnia Hospital District in Oulu, Finland. The current study included those who were alive and living in Finland in 1982 and 1997 and who answered a postal questionnaire sent to cohort members in 1997 (N = 8211), at the age of 31 years, with the exception of 84 individuals who had denied the use of their data.
2.2. Medication data Information on the use of AEDs and BZDs was received via a postal questionnaire sent to all cohort members in 1997–1998. This questionnaire included questions about the current use of prescribed medication, including the name and daily dose of any drugs used. All drug uses in our study population were classified according to the Anatomical Therapeutic Chemical (ATC) classification system [18], of which we selected groups N03A (antiepileptics), N05BA (anxiolytics, benzodiazepine derivatives), N05CD (hypnotics and sedatives, benzodiazepine derivatives), and N05CF (hypnotics and sedatives, benzodiazepine-related drugs) for this study. The group N03AE (benzodiazepine derivatives of the antiepileptic group) was included in both analyses of AEDs and BZDs. The AEDs used by our sample were divided into the following subgroups based on their WHO classification: N03AA (barbiturates and derivatives), N03AB (hydantoin derivatives), N03AE (benzodiazepine derivatives), N03AF (carboxamide derivatives), and N03AG (fatty acid derivatives). The BZDs were divided into subgroups in accordance with the WHO codes mentioned above. The subjects were also divided into two groups, depending on whether they used only one or more than one different drug at a time. In addition, the concomitant use of antidepressants or antipsychotics was studied. Doses of AEDs and BZDs were studied using the defined daily dose (DDD) classification. The DDD is the average maintenance dose estimated by the WHO based on global health statistics [18]. We divided the daily dose of the subject by the DDD in order to form the DDD ratio. In the cases for whom daily doses of BZDs were reported as a range of doses, the maximum daily dose of the subject was used. Subjects were classified into low dose and high dose categories by the median value of the DDD ratio. Individuals receiving AEDs or BZDs for whom no information on the daily dose was available (AEDs n = 4, BZDs n = 3) were excluded from the dose analyses.
199
2.3. Suicidal ideation and other depression and anxiety symptoms Suicidal ideation data were collected using the same postal questionnaire that assessed the use of medication [19]. The questionnaire included a 25-item version of the Hopkins Symptom Checklist (SCL) validated in Finnish [20], with questions on symptoms related to depression and anxiety experienced in the previous week. One question was “To what extent have you suffered from suicidal thoughts during the last week?” The answers were scored on a scale of 1 (“not at all”) to 4 (“very much”). We adjusted the score for suicidal ideation by other symptoms in the SCL by dividing the score for suicidal ideation by the mean score for other answers in the questionnaire (to form the SCL ratio) [21]. This adjustment was based on the indication in previous studies that suicidality correlates with the severity of the symptoms in the case of both depression and anxiety [22–24]. We also report the total mean score (without suicidal ideation) in the SCL (SCL total). Of all SCL questions, 13 items were classified as measuring depression symptoms, and 10 were classified as measuring anxiety symptoms. Those who left five or more items blank (n = 6) in the SCL were excluded from the series. 2.4. Definition of epilepsy All subjects with a diagnosis of epilepsy or seizures (i.e., ICD-8 345 and 331.2; ICD-9 345; ICD-10 G40–G41) during the period 1966–2004 were identified from the National Finnish Hospital Discharge Register and from the records of the Social Insurance Institution of Finland [25]. We also used data collected from cohort members or their parents, as prescribed by Löfgren et al. [25]. The patient files of all subjects predisposed to epilepsy were examined using systematic medical chart abstraction, and epilepsy was diagnosed if there had been at least two unprovoked seizures or one unprovoked seizure with EEG findings consistent with epilepsy, as defined by the International League Against Epilepsy [25,26]. The final epilepsy diagnoses were done by specially trained medical doctors, who, in problematic cases, consulted experienced neurologists. Only those individuals who were diagnosed with epilepsy before the end of 1997 were included in the analyses. 2.5. Psychiatric diagnoses and symptoms Information on psychiatric disorder diagnoses was collected from both nationwide registers and a 31-year follow-up questionnaire. All cohort members over 16 years of age, who had appeared in the Finnish Hospital Discharge Register (FHDR) up to 1997 in association with any psychiatric disorder (i.e., ICD-8 290–309; ICD-9 diagnoses 290–316; and ICD-10 F00–F69 and F99), were identified. All case records until 1997 were scrutinized, and diagnoses were validated for the DSM-III-R criteria, as prescribed in detail by Isohanni et al. and Moilanen et al. [27,28]. The validation was done by six psychiatrists or psychiatry trainees; all cases were first examined by two evaluators independently, and, in problematic cases, the diagnoses were done in a consensus panel including experts in psychiatric diagnostics. The good reliability of raters of psychiatric diagnoses has been shown in the paper of Moilanen et al. [28]. The validated diagnoses were combined with register information from the Social Insurance Institution of Finland on disability pensions, sick days, and reimbursement-eligible medicines for psychiatric disorders. Self-reported diagnoses were collected from a postal questionnaire sent to all subjects in 1997, asking whether the subject had had any of the following conditions diagnosed by a medical doctor: depression, psychosis, an alcohol-use disorder, some other substance use disorder, or any other psychiatric disorder. All these data were combined, excluding those who had organic mental disorder, e.g., acute intoxication or delirium caused by different substances or medical conditions (ICD-8 codes 290–294 and 309; ICD-9 290–294 and 310; and ICD-10 F00–F09, F1x.0, and F1x.3–F1x.9).
200
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
2.6. Attrition analysis The cohort included 10,799 subjects (5487 (50.8%) males) alive and living in Finland in 1997. Attrition and the validity of the questionnaire data have been described in detail by Haapea et al. [17,19]. Of all of the cohort members, 8211 (76.0%) answered the postal questionnaire. Females were more active in responding than males (80.6% vs. 71.7%). 2.7. Statistical analyses We present mean values (and standard deviations, SDs) for the SCL variable suicidal ideation, SCL total, SCL ratio, and depression and anxiety subscale scores for subjects with medication and those without medication. We also present mean values and SDs of the DDD ratio. Differences in continuous variables between AED and BZD users and nonusers were tested using Student's t-test and one-way ANOVA. The variable suicidal ideation (Likert-type variable with 4 categories) was tested using the Mann–Whitney U-test and the Kruskal–Wallis test; also, in these analyses, means and SDs are presented as the medians, and interquartile ranges were not presentative (values were mainly one). We used IBM SPSS 21 for statistical analyses. 3. Results The entire sample consisted of 8211 subjects who answered the postal questionnaire in 1997. Of these, 132 (1.6%) subjects were diagnosed as having epilepsy, 655 (8.0%) as having any nonorganic mental disorder, 372 (4.5%) as having depression, and 147 (1.8%) as having any substance use disorder. 3.1. Use of AEDs According to the postal questionnaire, AEDs were used by 54 (0.7%) subjects, of which 2 (3.7%) used phenobarbital (N03AA02), 2 (3.7%) used phenytoin (N03AB02), 8 (14.8%) used clonazepam (N03AE01), 26 (48.1%) used carbamazepine (N03AF01), 7 (13.0%) used oxcarbazepine (N03AF02), 15 (27.8%) used valproic acid (N03AG01), and 2 (3.7%) used vigabatrin (N03AG04). Fifty (92.6%) subjects used AEDs as monotherapy, and four (7.4%) subjects used AEDs as polytherapy. The mean dose (n = 50) of AEDs divided by the defined daily dose (DDD ratio) was 0.68 (standard deviation, SD = 0.51). The mean DDD was 0.74 (SD = 0.52) among those diagnosed with epilepsy and 0.32 (SD =
0.27) among those without an epilepsy diagnosis. Of the 54 AED users, 43 (79.6%) had epilepsy, whereas 13 (24.1%) had any nonorganic mental disorder, and 5 (9.3%) had neither of these. 3.2. Use of BZDs Benzodiazepines were used by 147 (1.8%) subjects. Only one BZD was used by 131 (89.1%) subjects and two or more drugs by 16 (10.9%) subjects. The BZD subgroup N05BA (diazepam, chlordiazepoxide, oxazepam, potassium clorazepate, lorazepam, or alprazolam) was used by 82 (59.0%) subjects, N05CD (nitrazepam, temazepam, or midazolam) by 34 (24.5%) subjects, N03AE (clonazepam) by 8 (5.4%) subjects, and N05CF (zopiclone or zolpidem) by 39 (28.1%) subjects. The mean dose (n = 144) of BZDs divided by the recommended daily dose (DDD ratio) was 1.06 (standard deviation, SD = 0.90). Of the 147 subjects using BZDs, 90 (61.2%) had any nonorganic mental disorder, 7 (4.8%) had epilepsy, and 54 (36.7%) had neither of these. 3.3. AEDs and suicidal ideation Overall, subjects using AEDs did not have higher suicidal ideation compared with the control population (Table 1). However, they had statistically significantly more other symptoms of depression and anxiety (SCL total score) (p = 0.004), and, when the figures were adjusted for suicidal ideation with other symptoms (SCL ratio), they experienced statistically significantly less suicidal ideation compared with the control population (p = 0.002). Among subjects with lifetime epilepsy, there were no statistically significant differences in suicidal ideation or symptoms of depression and anxiety between subjects who were using AEDs and subjects who were not using them (Table 1). Among subjects without epilepsy, there were no statistically significant differences in suicidal ideation, but subjects using AEDs for indications other than epilepsy had statistically significantly more other symptoms of depression and anxiety compared with subjects who were not using AEDs (p b 0.001). Among subjects with any nonorganic mental disorder, there were no statistically significant differences in suicidal ideation or symptoms of depression and anxiety between subjects using AEDs and subjects not using them (Table 1). There were no statistically significant differences in suicidal ideation between different types or doses of AEDs, AED polypharmacy, or those with or without adjunctive antidepressant or antipsychotic medication
Table 1 Suicidal ideation and the use of antiepileptic drugs (AEDs) among different diagnostic groups. Use of AEDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
SCL depression mean (SD)
SCL anxiety mean (SD)
No (n = 8157) Yes (n = 54)
1.05 (0.26) 1.07 (0.33)
1.35 (0.32)* 1.51 (0.40)*
0.80 (0.16)* 0.73 (0.17)*
1.39 (0.39)* 1.55 (0.47)*
1.31 (0.31)* 1.48 (0.41)*
No (n = 89) Yes (n = 43) No (n = 8068) Yes (n = 11)
1.13 (0.46) 1.05 (0.21) 1.05 (0.26) 1.18 (0.60)
1.39 (0.40) 1.42 (0.35) 1.35 (0.32)** 1.88 (0.38)**
0.83 (0.20) 0.76 (0.15) 0.80 (0.16)** 0.62 (0.19)**
1.43 (0.46) 1.47 (0.43) 1.39 (0.39)** 1.85 (0.50)**
1.36 (0.39) 1.38 (0.36) 1.31 (0.31)** 1.88 (0.34)**
Any nonorganic mental disorder Yes (n = 655) No (n = 642) Yes (n = 13) No (n = 7556) No (n = 7515) Yes (n = 41)
1.26 (0.61) 1.31 (0.63) 1.03 (0.20) 1.00 (0.00)
1.69 (0.49) 1.89 (0.51) 1.32 (0.28) 1.39 (0.27)
0.75 (0.24) 0.70 (0.23) 0.80 (0.15)* 0.74 (0.14)*
1.76 (0.56) 1.96 (0.59) 1.36 (0.35) 1.42 (0.33)
1.62 (0.51) 1.80 (0.51) 1.28 (0.27) 1.38 (0.32)
No epilepsy, no mental disorder No (n = 7437) Yes (n = 5)
1.03 (0.20) 1.00 (0.00)
1.32 (0.28)* 1.62 (0.21)*
0.80 (0.15)* 0.63 (0.08)*
1.36 (0.35) 1.48 (0.34)
1.28 (0.27)** 1.74 (0.21)**
All subjects N = 8211
Epilepsy Yes (n = 132) No (n = 8079)
Differences between AED users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for SCL total score without suicidal ideation.
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
(Table 2). Those using both antiepileptic medication and benzodiazepine medication had statistically significantly more suicidal ideation compared with those using only antiepileptic medication (p = 0.002), and this finding remained after adjusting for other symptoms of depression and anxiety (SCL ratio) (p = 0.03). However, there were only six subjects using concomitant antiepileptics and benzodiazepines.
3.4. BZDs and suicidality In overall terms, subjects using BZDs had higher suicidal ideation compared with the control population (Table 3). However, they also had more other symptoms of depression and anxiety and, after adjusting the suicidal ideation for other symptoms (SCL ratio), had less suicidality compared with the controls. All of these findings were statistically significant (p b 0.001). There was no statistically significant difference (p = 0.052) in suicidal ideation between BZD users and nonusers in subjects with epilepsy, although this relates to the small number (n = 7) of patients with epilepsy using BZDs. Among subjects without epilepsy, the results were similar to those of the whole sample, e.g., suicidal ideation was more common among BZD users (p b 0.001). Among subjects with any nonorganic mental disorder, the use of BZDs was associated with more suicidal ideation (p b 0.001) (Table 3). However, this association diminished after adjusting for other symptoms of depression and anxiety. Among subjects with no diagnosed mental disorder, there was no statistically significant difference in suicidal ideation between subjects using BZDs and subjects not using them. However, subjects using BZDs had more other symptoms of depression and anxiety, and, when suicidal ideation was adjusted for other symptoms (SCL ratio), the subjects using BZDs showed less suicidality compared with the controls.
Table 2 Suicidal ideation and the use of AEDs in different medication groups. Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
Class of AEDs N03AE (n = 8) N03AF (n = 33) N03AG (n = 17)
1.13 (0.35) 1.09 (0.38) 1.00 (0.00)
1.80 (0.51)* 1.43 (0.36) 1.46 (0.34)
0.64 (0.14) 0.77 (0.16)* 0.72 (0.16)
Number of AEDs One (n = 50) More than one (n = 4)
1.08 (0.34) 1.00 (0.00)
1.53 (0.40) 1.23 (0.12)
0.73 (0.17) 0.82 (0.08)
Dose of AEDsc Low dose (n = 18) Middle dose (n = 16) High dose (n = 15)
1.17 (0.51) 1.00 (0.00) 1.07 (0.26)
1.65 (0.48) 1.35 (0.31) 1.46 (0.32)
0.71 (0.18) 0.77 (0.16) 0.75 (0.15)
Concomitant use of BZDs BZD (n = 6) No BZD (n = 48)
1.50 (0.84)* 1.02 (0.14)*
1.68 (0.61) 1.49 (0.37)
0.87 (0.21)* 0.72 (0.15)*
Concomitant use of antidepressants Antidepressant (n = 4) 1.50 (1.00) No antidepressant (n = 50) 1.04 (0.20)
1.89 (0.55) 1.48 (0.38)
0.75 (0.28) 0.73 (0.16)
Concomitant use of antipsychotics Antipsychotics (n = 4) 1.50 (1.00) No antipsychotics (n = 50) 1.04 (0.20)
1.89 (0.55) 1.48 (0.38)
0.75 (0.28) 0.73 (0.16)
Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 when compared with other AEDs. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for SCL total score without suicidal ideation. c Subjects were classified into low dose, middle dose, and high dose categories based on the tertile values (0.40 and 0.60) of the defined daily dose (DDD) ratio. The Kruskal– Wallis test for suicidal ideation and one-way ANOVA for other variables. N03AE = benzodiazepine derivatives, N03AF = carboxamide derivatives, N03AG = fatty acid derivatives. N03AA (barbiturates and derivatives) and N03AB (hydantoin derivatives) were not presented because of small sample sizes (in both n = 2).
201
There were no statistically significant differences in suicidal ideation between different types or doses of BZDs or BZD polypharmacy (Table 4). However, subjects using both BZDs and concomitant antidepressant or antipsychotic medication had statistically significantly more suicidal ideation compared with subjects not using concomitant medications, but this difference diminished after adjusting for other symptoms of depression and anxiety. Additionally, we studied suicidal ideation and BZD use separately among those with high, middle, or low measured levels of insomnia, anxiety, and depression (Table 5). Among those who scored low or middle for insomnia, BZD use was significantly associated with higher suicidal ideation. However, when other symptoms were taken into account, BZD use was associated with less suicidality among those with low insomnia, and there were no differences among those with middle insomnia. Among those who scored high for insomnia, there was no statistically significant difference in suicidal ideation between BZD users and nonusers. Among those with both high and low anxiety symptoms, BZD use was associated with increased suicidal ideation. However, when adjusted for other symptoms, the association disappeared among those with high symptom levels, and, among those with low symptom levels, BZD use was associated with lower adjusted suicidal ideation. Among those with high scores of depression, the use of BZDs was associated with higher suicidal ideation, but this disappeared after adjustment.
3.5. Sensitivity analyses As a sensitivity analysis, ordinal regression analysis was performed to verify the statistically significant bivariate comparisons of suicidal ideation. The results of sensitivity analyses were similar to those of original analyses. We used the SCL total score as a covariate to replace the SCL ratio in adjusting for other symptoms of depression and anxiety. In all other analyses, the association between use of AED or BZD medications and increased suicidal ideation was no longer statistically significant after adjusting for other symptoms, except that, among subjects using BZDs, those using concomitant antipsychotics had more suicidal ideation compared with those not using concomitant antipsychotics (p = 0.038). We also performed ordinal regression to study the impact of the use of BZDs and insomnia, anxiety, or depression on suicidal ideation. With insomnia and depression, the use of BZDs was statistically significantly related to increased suicidal ideation (insomnia p b 0.001, depression p = 0.001), but, when adding the SCL total score to the top model as a covariate, the associations disappeared. With anxiety, the use of BZDs was not statistically significantly associated with suicidal ideation (p = 0.058).
4. Discussion 4.1. Main results In our study of the use of AEDs and suicidal ideation in all subjects, in subjects with epilepsy, or in subjects without epilepsy, we found no statistically significant differences between subjects using the medication and subjects not using the medication. When adjusting for the severity of depression and anxiety symptoms, subjects using AEDs had significantly less suicidal ideation compared with nonusers in all subjects and in subjects without epilepsy. These findings suggest that AEDs are not associated with an increased risk of suicidal ideation. Subjects using BZDs had statistically significantly more suicidal ideation compared with nonusers. However, when other symptoms of depression and anxiety were taken into account, subjects using BZDs had less suicidal ideation compared with nonusers. This finding shows that subjects with suicidal ideation have even more other symptoms of depression and anxiety, and the medication had no specific association with increased suicidal ideation.
202
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
Table 3 Suicidal ideation and the use of benzodiazepines (BZDs) in different diagnostic groups. BZDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
SCL depression mean (SD)
SCL anxiety mean (SD)
No (n = 8064) Yes (n = 147)
1.05 (0.25)** 1.29 (0.60)**
1.34 (0.31)** 1.80 (0.49)**
0.80 (0.16)** 0.72 (0.22)**
1.39 (0.38)** 1.85 (0.56)**
1.30 (0.30)** 1.76 (0.52)**
No (n = 125) Yes (n = 7) No (n = 7939) Yes (n = 140)
1.10 (0.39) 1.29 (0.49) 1.05 (0.25)** 1.29 (0.61)**
1.40 (0.37) 1.48 (0.61) 1.34 (0.31)** 1.82 (0.48)**
0.80 (0.19) 0.88 (0.09) 0.80 (0.16)** 0.71 (0.22)**
1.44 (0.43) 1.54 (0.76) 1.39 (0.38)** 1.86 (0.54)**
1.36 (0.37) 1.44 (0.57) 1.30 (0.30)** 1.77 (0.52)**
Any nonorganic mental disorder (N = 8211) Yes (n = 655) No (n = 565) Yes (n = 90) No (n = 7556) No (n = 7499) Yes (n = 57)
1.24 (0.59)** 1.43 (0.70)** 1.03 (0.19) 1.07 (0.26)
1.65 (0.48)** 1.96 (0.48)** 1.32 (0.28)** 1.56 (0.40)**
0.76 (0.23) 0.72 (0.24) 0.80 (0.15)** 0.72 (0.18)**
1.73 (0.55)** 2.02 (0.55)** 1.36 (0.35)** 1.57 (0.45)**
1.58 (0.49)** 1.90 (0.53)** 1.28 (0.27)** 1.53 (0.43)**
Any depression (N = 8211) Yes (n = 372) No (n = 308) Yes (n = 64) No (n = 7839) No (n = 7756) Yes (n = 83)
1.27 (0.64)* 1.47 (0.69)* 1.04 (0.22)** 1.16 (0.48)**
1.69 (0.50)** 1.98 (0.46)** 1.33 (0.29)** 1.67 (0.47)**
0.76 (0.25) 0.74 (0.26) 0.80 (0.15)** 0.71 (0.18)**
1.78 (0.58)* 2.04 (0.54)* 1.37 (0.36)** 1.71 (0.53)**
1.61 (0.50)** 1.91 (0.50)** 1.29 (0.28)** 1.64 (0.51)**
Any substance use disorder (N = 8211) Yes (n = 147) No (n = 127) Yes (n = 20) No (n = 8064) No (n = 7937) Yes (n = 127)
1.43 (0.79) 1.60 (0.94) 1.04 (0.23)** 1.24 (0.52)**
1.78 (0.53)* 2.13 (0.56)* 1.33 (0.30)** 1.75 (0.46)**
0.79 (0.30) 0.73 (0.28) 0.80 (0.15)** 0.72 (0.21)**
1.83 (0.58)* 2.20 (0.66)* 1.38 (0.37)** 1.79 (0.52)**
1.71 (0.56)* 2.08 (0.60)* 1.30 (0.29)** 1.71 (0.49)**
All subjects (N = 8211)
Epilepsy (N = 8211) Yes (n = 132) No (n = 8079)
Differences between BZD users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation.
4.2. AEDs and suicidal ideation and behavior In previous studies, the use of AEDs has been associated with an increased risk of suicidal ideation and behavior (suicides and attempted suicides). In 2008, the FDA mandated a black box warning on AEDs
Table 4 Suicidal ideation and the use of BZDs in different medication groups. Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
Class of BZDs N03AE n = 8 N05BA n = 82 N05CD n = 34 N05CF n = 39
1.13 (0.35) 1.30 (0.62) 1.24 (0.50) 1.38 (0.67)
1.80 (0.51) 1.87 (0.47) 1.75 (0.46) 1.80 (0.54)
0.64 (0.14) 0.70 (0.22) 0.72 (0.21) 0.77 (0.25)
Number of BZDs One (n = 131) More than one (n = 16)
1.28 (0.60) 1.38 (0.62)
1.78 (0.50) 1.98 (0.36)
0.73 (0.22) 0.69 (0.23)
Dose of BZDsc Low dose (n = 47) Middle dose (n = 53) High dose (n = 36)
1.23 (0.52) 1.34 (0.73) 1.33 (0.54)
1.78 (0.44) 1.78 (0.59) 1.84 (0.40)
0.70 (0.20) 0.75 (0.22) 0.73 (0.26)
Concomitant use of antidepressant Antidepressant (n = 41) 1.44 (0.67)* No antidepressant (n = 106) 1.24 (0.56)*
1.98 (0.48)* 1.74 (0.48)*
0.72 (0.24) 0.72 (0.21)
Concomitant use of antipsychotics Antipsychotics (n = 21) 1.81 (0.93)** No antipsychotics (n = 126) 1.21 (0.48)**
2.18 (0.50)** 1.74 (0.46)**
0.79 (0.26) 0.71 (0.22)
Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 when compared with other BZDs. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation. c Subjects were classified into low dose, middle dose, and high dose categories based on the tertile values (0.75 and 1.00) of the defined daily dose (DDD) ratio. The Kruskal–Wallis test for suicidal ideation and one-way ANOVA for other variables. N05BA = anxiolytics, benzodiazepine derivatives; N05CD = hypnotics and sedatives, benzodiazepine derivatives; N05CF = hypnotics and sedatives, benzodiazepine-related drugs.
because of the increased risk of suicidality [10]. This warning was based on a meta-analysis that showed an overall 1.8-fold risk of suicidal ideation or behavior in the study, which included randomized controlled trials of 11 AEDs. The risk of suicidal behavior was found to be higher than the risk of suicidal ideation (odds ratio (OR) = 2.92 vs. 1.45). Divided into subgroups by the indication of medication, the highest risk was revealed in the epilepsy indication subgroup (OR = 3.53) compared with psychiatric indications (OR = 1.51) or other indications (OR = 1.87). Since the FDA warning was issued, there has been broad discussion of the suicidality risk of AEDs. In our study, we found no statistically significant differences in suicidal ideation to be associated with the use of AEDs. Furthermore, when we took account of other symptoms of depression and anxiety, the mean of suicidal ideation was even lower among subjects using AEDs compared with nonusers. Our finding among people with epilepsy is consistent with Arana et al. [11], who found in their large population-based study that the use of AEDs was not associated with an increased risk of suicidalrelated events among patients with epilepsy or bipolar disorder. However, they found an association between the use of AEDs and higher suicidality in patients with depression and in those without epilepsy, depression or bipolar disorder. In our study, we did not find any differences in suicidal ideation between AED users and nonusers in any diagnostic groups. In a study of a population of elderly people, there were no statistically significant differences in suicide-related behaviors between subjects using AEDs for new-onset epilepsy and subjects using AEDs for other indications [29]. In a previous study, no statistically significant differences were found between those using AEDs and those not using AEDs or lithium when examining the use of AEDs and suicide attempt rates in a large cohort of patients with bipolar disorder [30]. This finding is similar to our findings. Additionally, in the same study, it was found that the rate of suicide attempts was higher before the initiation of AEDs than afterwards, suggesting that the medicine had a protective effect [30]. Other studies have found similar temporal trends [31]. In our study, the use of AEDs seemed to be associated with a decreased risk of suicidal
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
203
Table 5 Suicidal ideation and the use of BZDs in different symptom severity groups.
SCL insomnia High (score: 3–4) Middle (score: 2) Low (score: 1)
SCL anxiety High (score: ≥1.75) Middle (score: N1.55 and b1.75) Low (score: b1.55) SCL depression High (score: ≥1.75) Middle (score: N1.55 and b1.75) Low (score: ≤1.55)
BZDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
No (n = 471) Yes (n = 50) No (n = 2212) Yes (n = 52) No (n = 5375) Yes (n = 45)
1.23 (0.57) 1.36 (0.66) 1.07 (0.29)** 1.31 (0.54)** 1.02 (0.16)** 1.20 (0.59)**
1.79 (0.43)** 2.04 (0.47)** 1.46 (0.31)** 1.74 (0.42)** 1.25 (0.24)** 1.62 (0.49)**
0.69 (0.22) 0.66 (0.22) 0.75 (0.16) 0.75 (0.23) 0.83 (0.14)* 0.75 (0.20)*
No (n = 620) Yes (n = 63) No (n = 646) Yes (n = 21) No (n = 6798) Yes (n = 63)
1.34 (0.66)* 1.54 (0.78)* 1.09 (0.29) 1.19 (0.40) 1.02 (0.13)** 1.08 (0.27)**
2.05 (0.37)* 2.21 (0.42)* 1.66 (0.19) 1.72 (0.14) 1.25 (0.19)** 1.42 (0.25)**
0.65 (0.25) 0.68 (0.25) 0.66 (0.16) 0.69 (0.21) 0.83 (0.13)* 0.77 (0.18)*
No (n = 1364) Yes (n = 86) No (n = 742) Yes (n = 11) No (n = 5958) Yes (n = 50)
1.23 (0.53)** 1.49 (0.72)** 1.04 (0.21) 1.00 (0.00) 1.01 (0.07) 1.02 (0.14)
1.85 (0.33)** 2.10 (0.41)** 1.52 (0.11)** 1.64 (0.21)** 1.20 (0.14)** 1.33 (0.18)**
0.66 (0.23) 0.70 (0.26) 0.69 (0.14) 0.61 (0.46) 0.85 (0.11)** 0.78 (0.14)**
Differences between BZD users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation.
ideation when we adjusted the figures for other symptoms of depression and anxiety.
4.3. BZDs and suicidal ideation and behavior In previous studies, the use of BZDs has been associated with increased risk of suicidality [12,32]. These findings are most probably at least partly confounded by the indication of medication and the severity of psychiatric symptoms in individuals using BZDs. In our study, we found the use of BZDs to be associated with increased suicidal ideation in general, but this association disappeared when adjusted for the severity of depression and anxiety symptoms. Our results suggest that the use of BZDs is associated not only with increased suicidal ideation but also with the severity of other depression and anxiety symptoms. Supportively, in a large population-based cohort study with a followup of over ten years, the use of BZDs was found to be associated with an increased risk of severe anxiety and depression symptoms, even when the severity of baseline symptoms and potential confounders were controlled for [33]. In a meta-analysis study of alprazolam, used in the treatment of depression, neither the risk of the emergence nor that of the worsening of suicidal ideation was significantly different between alprazolam and a placebo [15]. Additionally, the rate of improvement of suicidal ideation was greater for alprazolam than for the placebo. This finding is contrary to our finding that the use of BZDs was associated with an increased mean score for suicidal ideation among subjects with diagnosed depression. However, when we adjusted for other symptoms of depression and anxiety, we did not find any statistically significant differences in suicidal ideation between BZD users and nonusers among patients with depression. In a previous study, concomitant use of BZDs and antidepressants was associated with a higher risk of attempted suicide compared with the risk associated with the use of BZDs or antidepressants in monotherapy [13]. In our study, we found that the concomitant use of BZDs and antidepressants or antipsychotics was associated with an increased risk of suicidal ideation, but this association lost its significance when adjusted for other symptoms of depression and anxiety. This relates to the fact that BZD users experienced significantly more depression and
anxiety symptoms. A study on polypharmacy among patients with schizophrenia found that BZD use combined with antipsychotic or antidepressant use was associated with higher mortality, including suicide mortality, compared with no BZD use [34]. In a large population-based cohort study, the use of BZDs was associated with increased self-reported sleeping difficulties, even when baseline demographics and clinical characteristics, including severity of sleeping difficulties, were controlled for [33]. In our study, BZD use was associated significantly with more suicidal ideation among those who scored low for insomnia. However, when other symptoms were taken into account, BZD use was associated with less suicidality. Among those with high insomnia rates, there was no difference in suicidal ideation between BZD users and nonusers.
4.4. Strengths and limitations The main strength of our study lies on our unselected populationbased data, which enabled us to study subjects using AEDs and BZDs and suicidal ideation in a real-world setting, regardless of the indication of medication. We were also able to adjust the suicidal ideation for the severity of depression and anxiety and study the associations between AED and BZD use and suicidal ideation in different diagnostic groups. Haapea et al. [19] have shown that the questionnaire data on medication are reliable and consistent with the register data on this same birth cohort. One main limitation is that our data were collected more than a decade ago, and many new drugs have been introduced since 1997. Therefore, our results cannot be generalized to newer antiepileptics. The precise purpose of the medication was not known in all cases. In addition, this was a cross-sectional study of a cohort population and therefore we could not study causal relationships between medication and suicidal ideation. Although suicidal ideation is a prerequisite of attempted or committed suicide, it can be considered a different phenomenon [35]. The present results are not generalizable for suicidal behavior as a whole. Our sample sizes were small in some analyses, e.g., the number of individuals suffering from bipolar disorder was limited, and we were not, therefore, able to study this diagnostic group separately from other psychiatric diagnoses. However, the total sample
204
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
size was very large, including over 8000 individuals. Because of the small sample sizes in some groups, the SCL ratio was selected as an adjustment method instead of using multivariate analyses (e.g., ordinal regression). It should be noted that several bivariate analyses were performed, so some of the statistical significances might have occurred by chance. However, most of the significances were p b 0.001, which may minimize this issue. Another limitation is that we had no detailed information on the classification of epilepsy or, e.g., seizure frequency in patients with epilepsy. However, the study by Hecimovic et al. [36] showed that suicidal ideation was not associated with the epilepsy type or severity in people with epilepsy. 5. Conclusions In our cohort sample, AED use was not associated with an increased risk of suicidal ideation, especially if the subject did not suffer from a mental disorder. Benzodiazepine use was associated with higher suicidal ideation. However, when other symptoms of depression and anxiety were taken into account, subjects using BZDs experienced less suicidal ideation than nonusers. These medications do not specifically seem to increase suicidal ideation. Acknowledgments The study was supported by the Academy of Finland (132071, 268336, 278286); the European Commission (EURO-BLCS, Framework 5 award QLG1-CT-2000-01643); NARSAD: the Brain and Behavior Research Fund; the Northern Finland Health Care Support Foundation; the Finnish Medical Association; the Finnish Psychiatry Research Foundation; the Sigrid Jusélius Foundation; the Jalmari and Rauha Ahokas Foundation; the Finnish Medical Society Duodecim, Oulu; and the US National Institute of Mental Health (NIMH) (5R01MH63706:02). Conflict of interest The authors report no conflict of interest. References [1] Ernst C, Mechawar N, Turecki G. Suicide neurobiology. Prog Neurobiol 2009;89: 315–33. [2] Bell GS, Gaitatzis A, Bell CL, Johnson AL, Sander JW. Suicide in people with epilepsy: how great is the risk? Epilepsia 2009;50:1933–42. [3] Jones JE, Hermann BP, Barry JJ, Gilliam FG, Kanner AM, Meador KJ. Rates and risk factors for suicide, suicidal ideation, and suicide attempts in chronic epilepsy. Epilepsy Behav 2003;4(Suppl. 3):S31–8. [4] Christensen J, Vestergaard M, Mortensen PB, Sidenius P, Agerbo E. Epilepsy and risk of suicide: a population-based case–control study. Lancet Neurol 2007;6:693–8. [5] Kessler RC, Borges G, Walters EE. Prevalence of and risk factors for lifetime suicide attempts in the national comorbidity survey. Arch Gen Psychiatry 1999;56:617–26. [6] Kessler RC, Berglund P, Borges G, Nock M, Wang PS. Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990–1992 to 2001–2003. JAMA 2005; 293:2487–95. [7] Nock MK, Borges G, Bromet EJ, et al. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br J Psychiatry 2008;192:98–105. [8] Tellez-Zenteno JF, Patten SB, Jette N, Williams J, Wiebe S. Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia 2007;48:2336–44. [9] Gaitatzis A, Carroll K, Majeed AW, Sander J. The epidemiology of the comorbidity of epilepsy in the general population. Epilepsia 2004;45:1613–22. [10] US Food and Drug Administration. Antiepileptic drugs and suicidality. Available at http://Www.fda.gov/downloads/drugs/DrugSafety/PostmarketDrugSafety InformationforPatientsandProviders/UCM192556.pdf; 2008.
[11] Arana A, Wentworth CE, Ayuso-Mateos JL, Arellano FM. Suicide-related events in patients treated with antiepileptic drugs. N Engl J Med 2010;363:542–51. [12] Mallon L, Broman J, Hetta J. Is usage of hypnotics associated with mortality? Sleep Med 2009;10:279–86. [13] Neutel CI, Patten SB. Risk of suicide attempts after benzodiazepine and/or antidepressant use. Ann Epidemiol 1997;7:568–74. [14] Brower KJ, McCammon RJ, Wojnar M, Ilgen MA, Wojnar J, Valenstein M. Prescription sleeping pills, insomnia, and suicidality in the national comorbidity survey replication. J Clin Psychiatry 2011;72:515–21. [15] Jonas JM, Hearron Jr AE. Alprazolam and suicidal ideation: a meta-analysis of controlled trials in the treatment of depression. J Clin Psychopharmacol 1996;16: 208–11. [16] Wang P, Brookhart A, Ulbricht C, Schneeweiss S. The pharmacoepidemiology of psychiatric medications. In: Tsuang M, Tohen M, Jones P, editors. Textbook in psychiatric epidemiology. 3rd ed. Wiley-Blackwell; 2011. p. 155–65. [17] Haapea M, Miettunen J, Laara E, et al. Non-participation in a field survey with respect to psychiatric disorders. Scand J Public Health 2008;36:728–36. [18] WHO Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2011. Norwegian Institute of Public Health; 2010(Oslo). [19] Haapea M, Miettunen J, Lindeman S, Joukamaa M, Koponen H. Agreement between self-reported and pharmacy data on medication use in the Northern Finland 1966 birth cohort. Int J Methods Psychiatr Res 2010;19:88–96. [20] Veijola J, Jokelainen J, Laksy K, et al. The Hopkins Symptom Checklist — 25 in screening DSM-III-R axis-I disorders. Nord J Psychiatry 2003;57:119–23. [21] Rissanen I, Jaaskeiainen E, Alaraisanen A, Isohanni M, Koponen H, Miettunen J. Use of antipsychotic medication and suicidality — the Northern Finland birth cohort 1966 study. Hum Psychopharmacol 2011;26:e158–9. [22] Norton PJ, Temple SR, Pettit JW. Suicidal ideation and anxiety disorders: elevated risk or artifact of comorbid depression? J Behav Ther Exp Psychiatry 2008;39: 515–25. [23] Suokas JT, Perala J, Suominen K, Saarni S, Lonnqvist J, Suvisaari JM. Epidemiology of suicide attempts among persons with psychotic disorder in the general population. Schizophr Res 2010;124:22–8. [24] Diefenbach GJ, Woolley SB, Goethe JW. The association between self-reported anxiety symptoms and suicidality. J Nerv Ment Dis 2009;197:92–7. [25] Lofgren E, Pouta A, von Wendt L, Tapanainen J, Isojarvi JI, Jarvelin MR. Epilepsy in the Northern Finland birth cohort 1966 with special reference to fertility. Epilepsy Behav 2009;14:102–7. [26] Proposal for revised classification of epilepsies and epileptic syndromes. Commission on classification and terminology of the International League Against Epilepsy. Epilepsia 1989;30:389–99. [27] Isohanni M, Makikyro T, Moring J, et al. A comparison of clinical and research DSMIII-R diagnoses of schizophrenia in a Finnish national birth cohort. Clinical and research diagnoses of schizophrenia. Soc Psychiatry Psychiatr Epidemiol 1997;32: 303–8. [28] Moilanen K, Veijola J, Laksy K, et al. Reasons for the diagnostic discordance between clinicians and researchers in schizophrenia in the Northern Finland 1966 birth cohort. Soc Psychiatry Psychiatr Epidemiol 2003;38:305–10. [29] VanCott AC, Cramer JA, Copeland LA, et al. Suicide-related behaviors in older patients with new anti-epileptic drug use: data from the VA hospital system. BMC Med 2010; 8:4. [30] Gibbons RD, Hur K, Brown CH, Mann JJ. Relationship between antiepileptic drugs and suicide attempts in patients with bipolar disorder. Arch Gen Psychiatry 2009; 66:1354–60. [31] Pugh MJ, Hesdorffer D, Wang CP, et al. Temporal trends in new exposure to antiepileptic drug monotherapy and suicide-related behavior. Neurology 2013;81:1900–6. [32] Voaklander DC, Rowe BH, Dryden DM, Pahal J, Saar P, Kelly KD. Medical illness, medication use and suicide in seniors: a population-based case–control study. J Epidemiol Community Health 2008;62:138–46. [33] Nordfjaern T. A population-based cohort study of anxiety, depression, sleep and alcohol outcomes among benzodiazepine and z-hypnotic users. Addict Behav 2012; 37:1151–7. [34] Tiihonen J, Suokas JT, Suvisaari JM, Haukka J, Korhonen P. Polypharmacy with antipsychotics, antidepressants, or benzodiazepines and mortality in schizophrenia. Arch Gen Psychiatry 2012;69:476–83. [35] O'Carroll PW, Berman AL, Maris RW, Moscicki EK, Tanney BL, Silverman MM. Beyond the tower of Babel: a nomenclature for suicidology. Suicide Life Threat Behav 1996; 26:237–52. [36] Hecimovic H, Santos JM, Carter J, et al. Depression but not seizure factors or quality of life predicts suicidality in epilepsy. Epilepsy Behav 2012;24:426–9.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Use of antiepileptic or benzodiazepine medication and suicidal ideation — The Northern Finland Birth Cohort 1966 I. Rissanen a,b,⁎, E. Jääskeläinen a,b,c, M. Isohanni a,d, H. Koponen e,f, H. Ansakorpi b,g, J. Miettunen a,b,c,d a
Department of Psychiatry, Institute of Clinical Medicine, P.O. Box 5000, 90014, University of Oulu, Oulu, Finland Medical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland University of Oulu, Institute of Health Sciences, P.O. Box 5000, FIN-90014, University of Oulu, Oulu, Finland d Department of Psychiatry, Oulu University Hospital, P.O. Box 26, FIN-90029, OYS, Oulu, Finland e Department of Psychiatry, Institute of Clinical Medicine, University of Helsinki, Helsinki, Finland f Department of Psychiatry, Helsinki University Central Hospital, Helsinki, Finland g Department of Neurology, Oulu University Hospital, University of Oulu, Oulu, Finland b c
a r t i c l e
i n f o
Article history: Received 31 October 2014 Revised 20 January 2015 Accepted 1 March 2015 Available online 29 April 2015 Keywords: Antiepileptic drugs Benzodiazepines Suicidal ideation Cohort study Depression Anxiety
a b s t r a c t Both antiepileptic drugs (AEDs) and benzodiazepines (BZDs) have previously been associated with an increased risk of suicidality. Our aim was to study the association between the use of conventional AEDs and BZDs and suicidal ideation in a large population-based cohort. Information on the medications used in the Northern Finland Birth Cohort 1966 was collected from the subjects at the age of 31 years, using a postal questionnaire (N = 8211). The presence of suicidal ideation and other symptoms of depression and anxiety was assessed via the Hopkins Symptom Checklist — 25 questionnaire. The associations between medications and suicidal ideation were studied in different diagnostic groups and adjusted for symptoms of depression and anxiety. No difference was observed in suicidal ideation between AED users (n = 54) and nonusers (n = 8157). Subjects using BZDs (n = 147) had greater suicidal ideation compared with nonusers (n = 8064). Antiepileptic drug and benzodiazepine users more often exhibited other depression and anxiety symptoms. After adjustment for these symptoms, both AED and BZD users had less suicidal ideation compared with nonusers. In conclusion, in this populationbased cohort, neither the use of AEDs nor that of BZDs was found to be associated with increased suicidal ideation when the symptoms of depression and anxiety were taken into account. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Suicidality is a severe medical condition related not only to several psychiatric disorders, including depression and anxiety [1], but also to epilepsy [2–4]. Suicidality can be divided into suicidal behavior (attempted or committed suicide) and suicidal ideation. Suicidal ideation is usually considered to be a risk factor for suicidal behavior [5–7]. A previous study found the lifetime prevalence of suicidal ideation in people with epilepsy to be 25% [8]. Patients with epilepsy also have a high comorbidity of psychiatric disorders [9]. A high rate of suicidal ideation and behavior among patients with epilepsy associates with comorbid psychiatric disorders, particularly depression and anxiety disorders [3,4]. In 2008, the United States Food and Drug Administration (FDA) mandated a black box warning on antiepileptic drugs (AEDs) because
⁎ Corresponding author at: Institute of Clinical Medicine, Department of Psychiatry, P.O. BOX 5000, 90014, University of Oulu, Oulu, Finland. E-mail address: [email protected].fi (I. Rissanen).
http://dx.doi.org/10.1016/j.yebeh.2015.03.001 1525-5050/© 2015 Elsevier Inc. All rights reserved.
of the increased risk of suicidality [10]. This warning was based on a meta-analysis that revealed a 1.8-fold risk of suicidal ideation or suicidal behavior in AEDs. The FDA study has been widely discussed, and the reliability of the results has been questioned. In contrast to the FDA study, a study by Arana et al. [11] suggested that AEDs were not associated with an increased risk of suicidality in people with epilepsy but were associated with an increased risk of suicidality in people with depression or in people without epilepsy, depression, or bipolar disorder. Benzodiazepine drugs (BZDs) not only are generally used as anxiolytics or hypnotics and sedatives but also have anticonvulsant properties, and some can be used as add-on therapy for epilepsy. Regular use of hypnotics has been associated with increased all-cause mortality and suicide mortality [12]. The use of BZDs has previously been associated with an increased risk of suicide attempts, especially among individuals who had not recently received antidepressant medication [13]. In addition, the use of sedative–hypnotic drugs has been associated with an increased risk of suicidal ideation, plans, and attempts, even when adjusted for demographic characteristics, physical and mental disorders, and sleep disturbances [14]. By contrast, in a meta-analysis studying the use of alprazolam (a benzodiazepine derivate) in the
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
treatment of depression, alprazolam bore a similar risk of the emergence and worsening of suicidal ideation as a placebo, but it showed a higher rate of improvement compared with the placebo in the case of suicidal ideation [15]. Previous studies of AEDs or BZDs and suicidal ideation have been unable to control their findings for other psychiatric symptoms and diagnoses. Few population-based studies have been performed on this subject, although such studies are optimal for estimating the adverse effects of medications [16]. Our aim was to study the associations between AEDs or BZDs and suicidal ideation among groups of different diagnoses or symptoms in a large population-based cohort. Our hypothesis was that both AEDs and BZDs relate not only to increased suicidal ideation but also to other symptoms of depression and anxiety.
2. Methods 2.1. The Northern Finland Birth Cohort 1966 The Northern Finland Birth Cohort 1966 (NFBC 1966) is a populationbased sample containing data on 12,058 babies born alive in the Finnish provinces of Oulu and Lapland with an expected date of birth in 1966 [17]. Since pregnancy, prospective data have been collected on, e.g., the cohort members' health, socioeconomic conditions, families, and living. Permission to gather data was obtained from the Ministry of Social and Health Affairs, and the study was approved by the Ethical Committee of the Northern Ostrobothnia Hospital District in Oulu, Finland. The current study included those who were alive and living in Finland in 1982 and 1997 and who answered a postal questionnaire sent to cohort members in 1997 (N = 8211), at the age of 31 years, with the exception of 84 individuals who had denied the use of their data.
2.2. Medication data Information on the use of AEDs and BZDs was received via a postal questionnaire sent to all cohort members in 1997–1998. This questionnaire included questions about the current use of prescribed medication, including the name and daily dose of any drugs used. All drug uses in our study population were classified according to the Anatomical Therapeutic Chemical (ATC) classification system [18], of which we selected groups N03A (antiepileptics), N05BA (anxiolytics, benzodiazepine derivatives), N05CD (hypnotics and sedatives, benzodiazepine derivatives), and N05CF (hypnotics and sedatives, benzodiazepine-related drugs) for this study. The group N03AE (benzodiazepine derivatives of the antiepileptic group) was included in both analyses of AEDs and BZDs. The AEDs used by our sample were divided into the following subgroups based on their WHO classification: N03AA (barbiturates and derivatives), N03AB (hydantoin derivatives), N03AE (benzodiazepine derivatives), N03AF (carboxamide derivatives), and N03AG (fatty acid derivatives). The BZDs were divided into subgroups in accordance with the WHO codes mentioned above. The subjects were also divided into two groups, depending on whether they used only one or more than one different drug at a time. In addition, the concomitant use of antidepressants or antipsychotics was studied. Doses of AEDs and BZDs were studied using the defined daily dose (DDD) classification. The DDD is the average maintenance dose estimated by the WHO based on global health statistics [18]. We divided the daily dose of the subject by the DDD in order to form the DDD ratio. In the cases for whom daily doses of BZDs were reported as a range of doses, the maximum daily dose of the subject was used. Subjects were classified into low dose and high dose categories by the median value of the DDD ratio. Individuals receiving AEDs or BZDs for whom no information on the daily dose was available (AEDs n = 4, BZDs n = 3) were excluded from the dose analyses.
199
2.3. Suicidal ideation and other depression and anxiety symptoms Suicidal ideation data were collected using the same postal questionnaire that assessed the use of medication [19]. The questionnaire included a 25-item version of the Hopkins Symptom Checklist (SCL) validated in Finnish [20], with questions on symptoms related to depression and anxiety experienced in the previous week. One question was “To what extent have you suffered from suicidal thoughts during the last week?” The answers were scored on a scale of 1 (“not at all”) to 4 (“very much”). We adjusted the score for suicidal ideation by other symptoms in the SCL by dividing the score for suicidal ideation by the mean score for other answers in the questionnaire (to form the SCL ratio) [21]. This adjustment was based on the indication in previous studies that suicidality correlates with the severity of the symptoms in the case of both depression and anxiety [22–24]. We also report the total mean score (without suicidal ideation) in the SCL (SCL total). Of all SCL questions, 13 items were classified as measuring depression symptoms, and 10 were classified as measuring anxiety symptoms. Those who left five or more items blank (n = 6) in the SCL were excluded from the series. 2.4. Definition of epilepsy All subjects with a diagnosis of epilepsy or seizures (i.e., ICD-8 345 and 331.2; ICD-9 345; ICD-10 G40–G41) during the period 1966–2004 were identified from the National Finnish Hospital Discharge Register and from the records of the Social Insurance Institution of Finland [25]. We also used data collected from cohort members or their parents, as prescribed by Löfgren et al. [25]. The patient files of all subjects predisposed to epilepsy were examined using systematic medical chart abstraction, and epilepsy was diagnosed if there had been at least two unprovoked seizures or one unprovoked seizure with EEG findings consistent with epilepsy, as defined by the International League Against Epilepsy [25,26]. The final epilepsy diagnoses were done by specially trained medical doctors, who, in problematic cases, consulted experienced neurologists. Only those individuals who were diagnosed with epilepsy before the end of 1997 were included in the analyses. 2.5. Psychiatric diagnoses and symptoms Information on psychiatric disorder diagnoses was collected from both nationwide registers and a 31-year follow-up questionnaire. All cohort members over 16 years of age, who had appeared in the Finnish Hospital Discharge Register (FHDR) up to 1997 in association with any psychiatric disorder (i.e., ICD-8 290–309; ICD-9 diagnoses 290–316; and ICD-10 F00–F69 and F99), were identified. All case records until 1997 were scrutinized, and diagnoses were validated for the DSM-III-R criteria, as prescribed in detail by Isohanni et al. and Moilanen et al. [27,28]. The validation was done by six psychiatrists or psychiatry trainees; all cases were first examined by two evaluators independently, and, in problematic cases, the diagnoses were done in a consensus panel including experts in psychiatric diagnostics. The good reliability of raters of psychiatric diagnoses has been shown in the paper of Moilanen et al. [28]. The validated diagnoses were combined with register information from the Social Insurance Institution of Finland on disability pensions, sick days, and reimbursement-eligible medicines for psychiatric disorders. Self-reported diagnoses were collected from a postal questionnaire sent to all subjects in 1997, asking whether the subject had had any of the following conditions diagnosed by a medical doctor: depression, psychosis, an alcohol-use disorder, some other substance use disorder, or any other psychiatric disorder. All these data were combined, excluding those who had organic mental disorder, e.g., acute intoxication or delirium caused by different substances or medical conditions (ICD-8 codes 290–294 and 309; ICD-9 290–294 and 310; and ICD-10 F00–F09, F1x.0, and F1x.3–F1x.9).
200
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
2.6. Attrition analysis The cohort included 10,799 subjects (5487 (50.8%) males) alive and living in Finland in 1997. Attrition and the validity of the questionnaire data have been described in detail by Haapea et al. [17,19]. Of all of the cohort members, 8211 (76.0%) answered the postal questionnaire. Females were more active in responding than males (80.6% vs. 71.7%). 2.7. Statistical analyses We present mean values (and standard deviations, SDs) for the SCL variable suicidal ideation, SCL total, SCL ratio, and depression and anxiety subscale scores for subjects with medication and those without medication. We also present mean values and SDs of the DDD ratio. Differences in continuous variables between AED and BZD users and nonusers were tested using Student's t-test and one-way ANOVA. The variable suicidal ideation (Likert-type variable with 4 categories) was tested using the Mann–Whitney U-test and the Kruskal–Wallis test; also, in these analyses, means and SDs are presented as the medians, and interquartile ranges were not presentative (values were mainly one). We used IBM SPSS 21 for statistical analyses. 3. Results The entire sample consisted of 8211 subjects who answered the postal questionnaire in 1997. Of these, 132 (1.6%) subjects were diagnosed as having epilepsy, 655 (8.0%) as having any nonorganic mental disorder, 372 (4.5%) as having depression, and 147 (1.8%) as having any substance use disorder. 3.1. Use of AEDs According to the postal questionnaire, AEDs were used by 54 (0.7%) subjects, of which 2 (3.7%) used phenobarbital (N03AA02), 2 (3.7%) used phenytoin (N03AB02), 8 (14.8%) used clonazepam (N03AE01), 26 (48.1%) used carbamazepine (N03AF01), 7 (13.0%) used oxcarbazepine (N03AF02), 15 (27.8%) used valproic acid (N03AG01), and 2 (3.7%) used vigabatrin (N03AG04). Fifty (92.6%) subjects used AEDs as monotherapy, and four (7.4%) subjects used AEDs as polytherapy. The mean dose (n = 50) of AEDs divided by the defined daily dose (DDD ratio) was 0.68 (standard deviation, SD = 0.51). The mean DDD was 0.74 (SD = 0.52) among those diagnosed with epilepsy and 0.32 (SD =
0.27) among those without an epilepsy diagnosis. Of the 54 AED users, 43 (79.6%) had epilepsy, whereas 13 (24.1%) had any nonorganic mental disorder, and 5 (9.3%) had neither of these. 3.2. Use of BZDs Benzodiazepines were used by 147 (1.8%) subjects. Only one BZD was used by 131 (89.1%) subjects and two or more drugs by 16 (10.9%) subjects. The BZD subgroup N05BA (diazepam, chlordiazepoxide, oxazepam, potassium clorazepate, lorazepam, or alprazolam) was used by 82 (59.0%) subjects, N05CD (nitrazepam, temazepam, or midazolam) by 34 (24.5%) subjects, N03AE (clonazepam) by 8 (5.4%) subjects, and N05CF (zopiclone or zolpidem) by 39 (28.1%) subjects. The mean dose (n = 144) of BZDs divided by the recommended daily dose (DDD ratio) was 1.06 (standard deviation, SD = 0.90). Of the 147 subjects using BZDs, 90 (61.2%) had any nonorganic mental disorder, 7 (4.8%) had epilepsy, and 54 (36.7%) had neither of these. 3.3. AEDs and suicidal ideation Overall, subjects using AEDs did not have higher suicidal ideation compared with the control population (Table 1). However, they had statistically significantly more other symptoms of depression and anxiety (SCL total score) (p = 0.004), and, when the figures were adjusted for suicidal ideation with other symptoms (SCL ratio), they experienced statistically significantly less suicidal ideation compared with the control population (p = 0.002). Among subjects with lifetime epilepsy, there were no statistically significant differences in suicidal ideation or symptoms of depression and anxiety between subjects who were using AEDs and subjects who were not using them (Table 1). Among subjects without epilepsy, there were no statistically significant differences in suicidal ideation, but subjects using AEDs for indications other than epilepsy had statistically significantly more other symptoms of depression and anxiety compared with subjects who were not using AEDs (p b 0.001). Among subjects with any nonorganic mental disorder, there were no statistically significant differences in suicidal ideation or symptoms of depression and anxiety between subjects using AEDs and subjects not using them (Table 1). There were no statistically significant differences in suicidal ideation between different types or doses of AEDs, AED polypharmacy, or those with or without adjunctive antidepressant or antipsychotic medication
Table 1 Suicidal ideation and the use of antiepileptic drugs (AEDs) among different diagnostic groups. Use of AEDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
SCL depression mean (SD)
SCL anxiety mean (SD)
No (n = 8157) Yes (n = 54)
1.05 (0.26) 1.07 (0.33)
1.35 (0.32)* 1.51 (0.40)*
0.80 (0.16)* 0.73 (0.17)*
1.39 (0.39)* 1.55 (0.47)*
1.31 (0.31)* 1.48 (0.41)*
No (n = 89) Yes (n = 43) No (n = 8068) Yes (n = 11)
1.13 (0.46) 1.05 (0.21) 1.05 (0.26) 1.18 (0.60)
1.39 (0.40) 1.42 (0.35) 1.35 (0.32)** 1.88 (0.38)**
0.83 (0.20) 0.76 (0.15) 0.80 (0.16)** 0.62 (0.19)**
1.43 (0.46) 1.47 (0.43) 1.39 (0.39)** 1.85 (0.50)**
1.36 (0.39) 1.38 (0.36) 1.31 (0.31)** 1.88 (0.34)**
Any nonorganic mental disorder Yes (n = 655) No (n = 642) Yes (n = 13) No (n = 7556) No (n = 7515) Yes (n = 41)
1.26 (0.61) 1.31 (0.63) 1.03 (0.20) 1.00 (0.00)
1.69 (0.49) 1.89 (0.51) 1.32 (0.28) 1.39 (0.27)
0.75 (0.24) 0.70 (0.23) 0.80 (0.15)* 0.74 (0.14)*
1.76 (0.56) 1.96 (0.59) 1.36 (0.35) 1.42 (0.33)
1.62 (0.51) 1.80 (0.51) 1.28 (0.27) 1.38 (0.32)
No epilepsy, no mental disorder No (n = 7437) Yes (n = 5)
1.03 (0.20) 1.00 (0.00)
1.32 (0.28)* 1.62 (0.21)*
0.80 (0.15)* 0.63 (0.08)*
1.36 (0.35) 1.48 (0.34)
1.28 (0.27)** 1.74 (0.21)**
All subjects N = 8211
Epilepsy Yes (n = 132) No (n = 8079)
Differences between AED users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for SCL total score without suicidal ideation.
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
(Table 2). Those using both antiepileptic medication and benzodiazepine medication had statistically significantly more suicidal ideation compared with those using only antiepileptic medication (p = 0.002), and this finding remained after adjusting for other symptoms of depression and anxiety (SCL ratio) (p = 0.03). However, there were only six subjects using concomitant antiepileptics and benzodiazepines.
3.4. BZDs and suicidality In overall terms, subjects using BZDs had higher suicidal ideation compared with the control population (Table 3). However, they also had more other symptoms of depression and anxiety and, after adjusting the suicidal ideation for other symptoms (SCL ratio), had less suicidality compared with the controls. All of these findings were statistically significant (p b 0.001). There was no statistically significant difference (p = 0.052) in suicidal ideation between BZD users and nonusers in subjects with epilepsy, although this relates to the small number (n = 7) of patients with epilepsy using BZDs. Among subjects without epilepsy, the results were similar to those of the whole sample, e.g., suicidal ideation was more common among BZD users (p b 0.001). Among subjects with any nonorganic mental disorder, the use of BZDs was associated with more suicidal ideation (p b 0.001) (Table 3). However, this association diminished after adjusting for other symptoms of depression and anxiety. Among subjects with no diagnosed mental disorder, there was no statistically significant difference in suicidal ideation between subjects using BZDs and subjects not using them. However, subjects using BZDs had more other symptoms of depression and anxiety, and, when suicidal ideation was adjusted for other symptoms (SCL ratio), the subjects using BZDs showed less suicidality compared with the controls.
Table 2 Suicidal ideation and the use of AEDs in different medication groups. Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
Class of AEDs N03AE (n = 8) N03AF (n = 33) N03AG (n = 17)
1.13 (0.35) 1.09 (0.38) 1.00 (0.00)
1.80 (0.51)* 1.43 (0.36) 1.46 (0.34)
0.64 (0.14) 0.77 (0.16)* 0.72 (0.16)
Number of AEDs One (n = 50) More than one (n = 4)
1.08 (0.34) 1.00 (0.00)
1.53 (0.40) 1.23 (0.12)
0.73 (0.17) 0.82 (0.08)
Dose of AEDsc Low dose (n = 18) Middle dose (n = 16) High dose (n = 15)
1.17 (0.51) 1.00 (0.00) 1.07 (0.26)
1.65 (0.48) 1.35 (0.31) 1.46 (0.32)
0.71 (0.18) 0.77 (0.16) 0.75 (0.15)
Concomitant use of BZDs BZD (n = 6) No BZD (n = 48)
1.50 (0.84)* 1.02 (0.14)*
1.68 (0.61) 1.49 (0.37)
0.87 (0.21)* 0.72 (0.15)*
Concomitant use of antidepressants Antidepressant (n = 4) 1.50 (1.00) No antidepressant (n = 50) 1.04 (0.20)
1.89 (0.55) 1.48 (0.38)
0.75 (0.28) 0.73 (0.16)
Concomitant use of antipsychotics Antipsychotics (n = 4) 1.50 (1.00) No antipsychotics (n = 50) 1.04 (0.20)
1.89 (0.55) 1.48 (0.38)
0.75 (0.28) 0.73 (0.16)
Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 when compared with other AEDs. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for SCL total score without suicidal ideation. c Subjects were classified into low dose, middle dose, and high dose categories based on the tertile values (0.40 and 0.60) of the defined daily dose (DDD) ratio. The Kruskal– Wallis test for suicidal ideation and one-way ANOVA for other variables. N03AE = benzodiazepine derivatives, N03AF = carboxamide derivatives, N03AG = fatty acid derivatives. N03AA (barbiturates and derivatives) and N03AB (hydantoin derivatives) were not presented because of small sample sizes (in both n = 2).
201
There were no statistically significant differences in suicidal ideation between different types or doses of BZDs or BZD polypharmacy (Table 4). However, subjects using both BZDs and concomitant antidepressant or antipsychotic medication had statistically significantly more suicidal ideation compared with subjects not using concomitant medications, but this difference diminished after adjusting for other symptoms of depression and anxiety. Additionally, we studied suicidal ideation and BZD use separately among those with high, middle, or low measured levels of insomnia, anxiety, and depression (Table 5). Among those who scored low or middle for insomnia, BZD use was significantly associated with higher suicidal ideation. However, when other symptoms were taken into account, BZD use was associated with less suicidality among those with low insomnia, and there were no differences among those with middle insomnia. Among those who scored high for insomnia, there was no statistically significant difference in suicidal ideation between BZD users and nonusers. Among those with both high and low anxiety symptoms, BZD use was associated with increased suicidal ideation. However, when adjusted for other symptoms, the association disappeared among those with high symptom levels, and, among those with low symptom levels, BZD use was associated with lower adjusted suicidal ideation. Among those with high scores of depression, the use of BZDs was associated with higher suicidal ideation, but this disappeared after adjustment.
3.5. Sensitivity analyses As a sensitivity analysis, ordinal regression analysis was performed to verify the statistically significant bivariate comparisons of suicidal ideation. The results of sensitivity analyses were similar to those of original analyses. We used the SCL total score as a covariate to replace the SCL ratio in adjusting for other symptoms of depression and anxiety. In all other analyses, the association between use of AED or BZD medications and increased suicidal ideation was no longer statistically significant after adjusting for other symptoms, except that, among subjects using BZDs, those using concomitant antipsychotics had more suicidal ideation compared with those not using concomitant antipsychotics (p = 0.038). We also performed ordinal regression to study the impact of the use of BZDs and insomnia, anxiety, or depression on suicidal ideation. With insomnia and depression, the use of BZDs was statistically significantly related to increased suicidal ideation (insomnia p b 0.001, depression p = 0.001), but, when adding the SCL total score to the top model as a covariate, the associations disappeared. With anxiety, the use of BZDs was not statistically significantly associated with suicidal ideation (p = 0.058).
4. Discussion 4.1. Main results In our study of the use of AEDs and suicidal ideation in all subjects, in subjects with epilepsy, or in subjects without epilepsy, we found no statistically significant differences between subjects using the medication and subjects not using the medication. When adjusting for the severity of depression and anxiety symptoms, subjects using AEDs had significantly less suicidal ideation compared with nonusers in all subjects and in subjects without epilepsy. These findings suggest that AEDs are not associated with an increased risk of suicidal ideation. Subjects using BZDs had statistically significantly more suicidal ideation compared with nonusers. However, when other symptoms of depression and anxiety were taken into account, subjects using BZDs had less suicidal ideation compared with nonusers. This finding shows that subjects with suicidal ideation have even more other symptoms of depression and anxiety, and the medication had no specific association with increased suicidal ideation.
202
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
Table 3 Suicidal ideation and the use of benzodiazepines (BZDs) in different diagnostic groups. BZDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
SCL depression mean (SD)
SCL anxiety mean (SD)
No (n = 8064) Yes (n = 147)
1.05 (0.25)** 1.29 (0.60)**
1.34 (0.31)** 1.80 (0.49)**
0.80 (0.16)** 0.72 (0.22)**
1.39 (0.38)** 1.85 (0.56)**
1.30 (0.30)** 1.76 (0.52)**
No (n = 125) Yes (n = 7) No (n = 7939) Yes (n = 140)
1.10 (0.39) 1.29 (0.49) 1.05 (0.25)** 1.29 (0.61)**
1.40 (0.37) 1.48 (0.61) 1.34 (0.31)** 1.82 (0.48)**
0.80 (0.19) 0.88 (0.09) 0.80 (0.16)** 0.71 (0.22)**
1.44 (0.43) 1.54 (0.76) 1.39 (0.38)** 1.86 (0.54)**
1.36 (0.37) 1.44 (0.57) 1.30 (0.30)** 1.77 (0.52)**
Any nonorganic mental disorder (N = 8211) Yes (n = 655) No (n = 565) Yes (n = 90) No (n = 7556) No (n = 7499) Yes (n = 57)
1.24 (0.59)** 1.43 (0.70)** 1.03 (0.19) 1.07 (0.26)
1.65 (0.48)** 1.96 (0.48)** 1.32 (0.28)** 1.56 (0.40)**
0.76 (0.23) 0.72 (0.24) 0.80 (0.15)** 0.72 (0.18)**
1.73 (0.55)** 2.02 (0.55)** 1.36 (0.35)** 1.57 (0.45)**
1.58 (0.49)** 1.90 (0.53)** 1.28 (0.27)** 1.53 (0.43)**
Any depression (N = 8211) Yes (n = 372) No (n = 308) Yes (n = 64) No (n = 7839) No (n = 7756) Yes (n = 83)
1.27 (0.64)* 1.47 (0.69)* 1.04 (0.22)** 1.16 (0.48)**
1.69 (0.50)** 1.98 (0.46)** 1.33 (0.29)** 1.67 (0.47)**
0.76 (0.25) 0.74 (0.26) 0.80 (0.15)** 0.71 (0.18)**
1.78 (0.58)* 2.04 (0.54)* 1.37 (0.36)** 1.71 (0.53)**
1.61 (0.50)** 1.91 (0.50)** 1.29 (0.28)** 1.64 (0.51)**
Any substance use disorder (N = 8211) Yes (n = 147) No (n = 127) Yes (n = 20) No (n = 8064) No (n = 7937) Yes (n = 127)
1.43 (0.79) 1.60 (0.94) 1.04 (0.23)** 1.24 (0.52)**
1.78 (0.53)* 2.13 (0.56)* 1.33 (0.30)** 1.75 (0.46)**
0.79 (0.30) 0.73 (0.28) 0.80 (0.15)** 0.72 (0.21)**
1.83 (0.58)* 2.20 (0.66)* 1.38 (0.37)** 1.79 (0.52)**
1.71 (0.56)* 2.08 (0.60)* 1.30 (0.29)** 1.71 (0.49)**
All subjects (N = 8211)
Epilepsy (N = 8211) Yes (n = 132) No (n = 8079)
Differences between BZD users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation.
4.2. AEDs and suicidal ideation and behavior In previous studies, the use of AEDs has been associated with an increased risk of suicidal ideation and behavior (suicides and attempted suicides). In 2008, the FDA mandated a black box warning on AEDs
Table 4 Suicidal ideation and the use of BZDs in different medication groups. Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
Class of BZDs N03AE n = 8 N05BA n = 82 N05CD n = 34 N05CF n = 39
1.13 (0.35) 1.30 (0.62) 1.24 (0.50) 1.38 (0.67)
1.80 (0.51) 1.87 (0.47) 1.75 (0.46) 1.80 (0.54)
0.64 (0.14) 0.70 (0.22) 0.72 (0.21) 0.77 (0.25)
Number of BZDs One (n = 131) More than one (n = 16)
1.28 (0.60) 1.38 (0.62)
1.78 (0.50) 1.98 (0.36)
0.73 (0.22) 0.69 (0.23)
Dose of BZDsc Low dose (n = 47) Middle dose (n = 53) High dose (n = 36)
1.23 (0.52) 1.34 (0.73) 1.33 (0.54)
1.78 (0.44) 1.78 (0.59) 1.84 (0.40)
0.70 (0.20) 0.75 (0.22) 0.73 (0.26)
Concomitant use of antidepressant Antidepressant (n = 41) 1.44 (0.67)* No antidepressant (n = 106) 1.24 (0.56)*
1.98 (0.48)* 1.74 (0.48)*
0.72 (0.24) 0.72 (0.21)
Concomitant use of antipsychotics Antipsychotics (n = 21) 1.81 (0.93)** No antipsychotics (n = 126) 1.21 (0.48)**
2.18 (0.50)** 1.74 (0.46)**
0.79 (0.26) 0.71 (0.22)
Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 when compared with other BZDs. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation. c Subjects were classified into low dose, middle dose, and high dose categories based on the tertile values (0.75 and 1.00) of the defined daily dose (DDD) ratio. The Kruskal–Wallis test for suicidal ideation and one-way ANOVA for other variables. N05BA = anxiolytics, benzodiazepine derivatives; N05CD = hypnotics and sedatives, benzodiazepine derivatives; N05CF = hypnotics and sedatives, benzodiazepine-related drugs.
because of the increased risk of suicidality [10]. This warning was based on a meta-analysis that showed an overall 1.8-fold risk of suicidal ideation or behavior in the study, which included randomized controlled trials of 11 AEDs. The risk of suicidal behavior was found to be higher than the risk of suicidal ideation (odds ratio (OR) = 2.92 vs. 1.45). Divided into subgroups by the indication of medication, the highest risk was revealed in the epilepsy indication subgroup (OR = 3.53) compared with psychiatric indications (OR = 1.51) or other indications (OR = 1.87). Since the FDA warning was issued, there has been broad discussion of the suicidality risk of AEDs. In our study, we found no statistically significant differences in suicidal ideation to be associated with the use of AEDs. Furthermore, when we took account of other symptoms of depression and anxiety, the mean of suicidal ideation was even lower among subjects using AEDs compared with nonusers. Our finding among people with epilepsy is consistent with Arana et al. [11], who found in their large population-based study that the use of AEDs was not associated with an increased risk of suicidalrelated events among patients with epilepsy or bipolar disorder. However, they found an association between the use of AEDs and higher suicidality in patients with depression and in those without epilepsy, depression or bipolar disorder. In our study, we did not find any differences in suicidal ideation between AED users and nonusers in any diagnostic groups. In a study of a population of elderly people, there were no statistically significant differences in suicide-related behaviors between subjects using AEDs for new-onset epilepsy and subjects using AEDs for other indications [29]. In a previous study, no statistically significant differences were found between those using AEDs and those not using AEDs or lithium when examining the use of AEDs and suicide attempt rates in a large cohort of patients with bipolar disorder [30]. This finding is similar to our findings. Additionally, in the same study, it was found that the rate of suicide attempts was higher before the initiation of AEDs than afterwards, suggesting that the medicine had a protective effect [30]. Other studies have found similar temporal trends [31]. In our study, the use of AEDs seemed to be associated with a decreased risk of suicidal
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
203
Table 5 Suicidal ideation and the use of BZDs in different symptom severity groups.
SCL insomnia High (score: 3–4) Middle (score: 2) Low (score: 1)
SCL anxiety High (score: ≥1.75) Middle (score: N1.55 and b1.75) Low (score: b1.55) SCL depression High (score: ≥1.75) Middle (score: N1.55 and b1.75) Low (score: ≤1.55)
BZDs
Suicidal ideation mean (SD)
SCL total scorea mean (SD)
SCL ratiob mean (SD)
No (n = 471) Yes (n = 50) No (n = 2212) Yes (n = 52) No (n = 5375) Yes (n = 45)
1.23 (0.57) 1.36 (0.66) 1.07 (0.29)** 1.31 (0.54)** 1.02 (0.16)** 1.20 (0.59)**
1.79 (0.43)** 2.04 (0.47)** 1.46 (0.31)** 1.74 (0.42)** 1.25 (0.24)** 1.62 (0.49)**
0.69 (0.22) 0.66 (0.22) 0.75 (0.16) 0.75 (0.23) 0.83 (0.14)* 0.75 (0.20)*
No (n = 620) Yes (n = 63) No (n = 646) Yes (n = 21) No (n = 6798) Yes (n = 63)
1.34 (0.66)* 1.54 (0.78)* 1.09 (0.29) 1.19 (0.40) 1.02 (0.13)** 1.08 (0.27)**
2.05 (0.37)* 2.21 (0.42)* 1.66 (0.19) 1.72 (0.14) 1.25 (0.19)** 1.42 (0.25)**
0.65 (0.25) 0.68 (0.25) 0.66 (0.16) 0.69 (0.21) 0.83 (0.13)* 0.77 (0.18)*
No (n = 1364) Yes (n = 86) No (n = 742) Yes (n = 11) No (n = 5958) Yes (n = 50)
1.23 (0.53)** 1.49 (0.72)** 1.04 (0.21) 1.00 (0.00) 1.01 (0.07) 1.02 (0.14)
1.85 (0.33)** 2.10 (0.41)** 1.52 (0.11)** 1.64 (0.21)** 1.20 (0.14)** 1.33 (0.18)**
0.66 (0.23) 0.70 (0.26) 0.69 (0.14) 0.61 (0.46) 0.85 (0.11)** 0.78 (0.14)**
Differences between BZD users and nonusers were tested using the Mann–Whitney U-test for suicidal ideation and Student's t-test for other variables, *p b 0.05 and **p b 0.001. SCL = Symptom Checklist — 25. a SCL total score without suicidal ideation. b Suicidal ideation adjusted for the SCL total score without suicidal ideation.
ideation when we adjusted the figures for other symptoms of depression and anxiety.
4.3. BZDs and suicidal ideation and behavior In previous studies, the use of BZDs has been associated with increased risk of suicidality [12,32]. These findings are most probably at least partly confounded by the indication of medication and the severity of psychiatric symptoms in individuals using BZDs. In our study, we found the use of BZDs to be associated with increased suicidal ideation in general, but this association disappeared when adjusted for the severity of depression and anxiety symptoms. Our results suggest that the use of BZDs is associated not only with increased suicidal ideation but also with the severity of other depression and anxiety symptoms. Supportively, in a large population-based cohort study with a followup of over ten years, the use of BZDs was found to be associated with an increased risk of severe anxiety and depression symptoms, even when the severity of baseline symptoms and potential confounders were controlled for [33]. In a meta-analysis study of alprazolam, used in the treatment of depression, neither the risk of the emergence nor that of the worsening of suicidal ideation was significantly different between alprazolam and a placebo [15]. Additionally, the rate of improvement of suicidal ideation was greater for alprazolam than for the placebo. This finding is contrary to our finding that the use of BZDs was associated with an increased mean score for suicidal ideation among subjects with diagnosed depression. However, when we adjusted for other symptoms of depression and anxiety, we did not find any statistically significant differences in suicidal ideation between BZD users and nonusers among patients with depression. In a previous study, concomitant use of BZDs and antidepressants was associated with a higher risk of attempted suicide compared with the risk associated with the use of BZDs or antidepressants in monotherapy [13]. In our study, we found that the concomitant use of BZDs and antidepressants or antipsychotics was associated with an increased risk of suicidal ideation, but this association lost its significance when adjusted for other symptoms of depression and anxiety. This relates to the fact that BZD users experienced significantly more depression and
anxiety symptoms. A study on polypharmacy among patients with schizophrenia found that BZD use combined with antipsychotic or antidepressant use was associated with higher mortality, including suicide mortality, compared with no BZD use [34]. In a large population-based cohort study, the use of BZDs was associated with increased self-reported sleeping difficulties, even when baseline demographics and clinical characteristics, including severity of sleeping difficulties, were controlled for [33]. In our study, BZD use was associated significantly with more suicidal ideation among those who scored low for insomnia. However, when other symptoms were taken into account, BZD use was associated with less suicidality. Among those with high insomnia rates, there was no difference in suicidal ideation between BZD users and nonusers.
4.4. Strengths and limitations The main strength of our study lies on our unselected populationbased data, which enabled us to study subjects using AEDs and BZDs and suicidal ideation in a real-world setting, regardless of the indication of medication. We were also able to adjust the suicidal ideation for the severity of depression and anxiety and study the associations between AED and BZD use and suicidal ideation in different diagnostic groups. Haapea et al. [19] have shown that the questionnaire data on medication are reliable and consistent with the register data on this same birth cohort. One main limitation is that our data were collected more than a decade ago, and many new drugs have been introduced since 1997. Therefore, our results cannot be generalized to newer antiepileptics. The precise purpose of the medication was not known in all cases. In addition, this was a cross-sectional study of a cohort population and therefore we could not study causal relationships between medication and suicidal ideation. Although suicidal ideation is a prerequisite of attempted or committed suicide, it can be considered a different phenomenon [35]. The present results are not generalizable for suicidal behavior as a whole. Our sample sizes were small in some analyses, e.g., the number of individuals suffering from bipolar disorder was limited, and we were not, therefore, able to study this diagnostic group separately from other psychiatric diagnoses. However, the total sample
204
I. Rissanen et al. / Epilepsy & Behavior 46 (2015) 198–204
size was very large, including over 8000 individuals. Because of the small sample sizes in some groups, the SCL ratio was selected as an adjustment method instead of using multivariate analyses (e.g., ordinal regression). It should be noted that several bivariate analyses were performed, so some of the statistical significances might have occurred by chance. However, most of the significances were p b 0.001, which may minimize this issue. Another limitation is that we had no detailed information on the classification of epilepsy or, e.g., seizure frequency in patients with epilepsy. However, the study by Hecimovic et al. [36] showed that suicidal ideation was not associated with the epilepsy type or severity in people with epilepsy. 5. Conclusions In our cohort sample, AED use was not associated with an increased risk of suicidal ideation, especially if the subject did not suffer from a mental disorder. Benzodiazepine use was associated with higher suicidal ideation. However, when other symptoms of depression and anxiety were taken into account, subjects using BZDs experienced less suicidal ideation than nonusers. These medications do not specifically seem to increase suicidal ideation. Acknowledgments The study was supported by the Academy of Finland (132071, 268336, 278286); the European Commission (EURO-BLCS, Framework 5 award QLG1-CT-2000-01643); NARSAD: the Brain and Behavior Research Fund; the Northern Finland Health Care Support Foundation; the Finnish Medical Association; the Finnish Psychiatry Research Foundation; the Sigrid Jusélius Foundation; the Jalmari and Rauha Ahokas Foundation; the Finnish Medical Society Duodecim, Oulu; and the US National Institute of Mental Health (NIMH) (5R01MH63706:02). Conflict of interest The authors report no conflict of interest. References [1] Ernst C, Mechawar N, Turecki G. Suicide neurobiology. Prog Neurobiol 2009;89: 315–33. [2] Bell GS, Gaitatzis A, Bell CL, Johnson AL, Sander JW. Suicide in people with epilepsy: how great is the risk? Epilepsia 2009;50:1933–42. [3] Jones JE, Hermann BP, Barry JJ, Gilliam FG, Kanner AM, Meador KJ. Rates and risk factors for suicide, suicidal ideation, and suicide attempts in chronic epilepsy. Epilepsy Behav 2003;4(Suppl. 3):S31–8. [4] Christensen J, Vestergaard M, Mortensen PB, Sidenius P, Agerbo E. Epilepsy and risk of suicide: a population-based case–control study. Lancet Neurol 2007;6:693–8. [5] Kessler RC, Borges G, Walters EE. Prevalence of and risk factors for lifetime suicide attempts in the national comorbidity survey. Arch Gen Psychiatry 1999;56:617–26. [6] Kessler RC, Berglund P, Borges G, Nock M, Wang PS. Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990–1992 to 2001–2003. JAMA 2005; 293:2487–95. [7] Nock MK, Borges G, Bromet EJ, et al. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br J Psychiatry 2008;192:98–105. [8] Tellez-Zenteno JF, Patten SB, Jette N, Williams J, Wiebe S. Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia 2007;48:2336–44. [9] Gaitatzis A, Carroll K, Majeed AW, Sander J. The epidemiology of the comorbidity of epilepsy in the general population. Epilepsia 2004;45:1613–22. [10] US Food and Drug Administration. Antiepileptic drugs and suicidality. Available at http://Www.fda.gov/downloads/drugs/DrugSafety/PostmarketDrugSafety InformationforPatientsandProviders/UCM192556.pdf; 2008.
[11] Arana A, Wentworth CE, Ayuso-Mateos JL, Arellano FM. Suicide-related events in patients treated with antiepileptic drugs. N Engl J Med 2010;363:542–51. [12] Mallon L, Broman J, Hetta J. Is usage of hypnotics associated with mortality? Sleep Med 2009;10:279–86. [13] Neutel CI, Patten SB. Risk of suicide attempts after benzodiazepine and/or antidepressant use. Ann Epidemiol 1997;7:568–74. [14] Brower KJ, McCammon RJ, Wojnar M, Ilgen MA, Wojnar J, Valenstein M. Prescription sleeping pills, insomnia, and suicidality in the national comorbidity survey replication. J Clin Psychiatry 2011;72:515–21. [15] Jonas JM, Hearron Jr AE. Alprazolam and suicidal ideation: a meta-analysis of controlled trials in the treatment of depression. J Clin Psychopharmacol 1996;16: 208–11. [16] Wang P, Brookhart A, Ulbricht C, Schneeweiss S. The pharmacoepidemiology of psychiatric medications. In: Tsuang M, Tohen M, Jones P, editors. Textbook in psychiatric epidemiology. 3rd ed. Wiley-Blackwell; 2011. p. 155–65. [17] Haapea M, Miettunen J, Laara E, et al. Non-participation in a field survey with respect to psychiatric disorders. Scand J Public Health 2008;36:728–36. [18] WHO Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2011. Norwegian Institute of Public Health; 2010(Oslo). [19] Haapea M, Miettunen J, Lindeman S, Joukamaa M, Koponen H. Agreement between self-reported and pharmacy data on medication use in the Northern Finland 1966 birth cohort. Int J Methods Psychiatr Res 2010;19:88–96. [20] Veijola J, Jokelainen J, Laksy K, et al. The Hopkins Symptom Checklist — 25 in screening DSM-III-R axis-I disorders. Nord J Psychiatry 2003;57:119–23. [21] Rissanen I, Jaaskeiainen E, Alaraisanen A, Isohanni M, Koponen H, Miettunen J. Use of antipsychotic medication and suicidality — the Northern Finland birth cohort 1966 study. Hum Psychopharmacol 2011;26:e158–9. [22] Norton PJ, Temple SR, Pettit JW. Suicidal ideation and anxiety disorders: elevated risk or artifact of comorbid depression? J Behav Ther Exp Psychiatry 2008;39: 515–25. [23] Suokas JT, Perala J, Suominen K, Saarni S, Lonnqvist J, Suvisaari JM. Epidemiology of suicide attempts among persons with psychotic disorder in the general population. Schizophr Res 2010;124:22–8. [24] Diefenbach GJ, Woolley SB, Goethe JW. The association between self-reported anxiety symptoms and suicidality. J Nerv Ment Dis 2009;197:92–7. [25] Lofgren E, Pouta A, von Wendt L, Tapanainen J, Isojarvi JI, Jarvelin MR. Epilepsy in the Northern Finland birth cohort 1966 with special reference to fertility. Epilepsy Behav 2009;14:102–7. [26] Proposal for revised classification of epilepsies and epileptic syndromes. Commission on classification and terminology of the International League Against Epilepsy. Epilepsia 1989;30:389–99. [27] Isohanni M, Makikyro T, Moring J, et al. A comparison of clinical and research DSMIII-R diagnoses of schizophrenia in a Finnish national birth cohort. Clinical and research diagnoses of schizophrenia. Soc Psychiatry Psychiatr Epidemiol 1997;32: 303–8. [28] Moilanen K, Veijola J, Laksy K, et al. Reasons for the diagnostic discordance between clinicians and researchers in schizophrenia in the Northern Finland 1966 birth cohort. Soc Psychiatry Psychiatr Epidemiol 2003;38:305–10. [29] VanCott AC, Cramer JA, Copeland LA, et al. Suicide-related behaviors in older patients with new anti-epileptic drug use: data from the VA hospital system. BMC Med 2010; 8:4. [30] Gibbons RD, Hur K, Brown CH, Mann JJ. Relationship between antiepileptic drugs and suicide attempts in patients with bipolar disorder. Arch Gen Psychiatry 2009; 66:1354–60. [31] Pugh MJ, Hesdorffer D, Wang CP, et al. Temporal trends in new exposure to antiepileptic drug monotherapy and suicide-related behavior. Neurology 2013;81:1900–6. [32] Voaklander DC, Rowe BH, Dryden DM, Pahal J, Saar P, Kelly KD. Medical illness, medication use and suicide in seniors: a population-based case–control study. J Epidemiol Community Health 2008;62:138–46. [33] Nordfjaern T. A population-based cohort study of anxiety, depression, sleep and alcohol outcomes among benzodiazepine and z-hypnotic users. Addict Behav 2012; 37:1151–7. [34] Tiihonen J, Suokas JT, Suvisaari JM, Haukka J, Korhonen P. Polypharmacy with antipsychotics, antidepressants, or benzodiazepines and mortality in schizophrenia. Arch Gen Psychiatry 2012;69:476–83. [35] O'Carroll PW, Berman AL, Maris RW, Moscicki EK, Tanney BL, Silverman MM. Beyond the tower of Babel: a nomenclature for suicidology. Suicide Life Threat Behav 1996; 26:237–52. [36] Hecimovic H, Santos JM, Carter J, et al. Depression but not seizure factors or quality of life predicts suicidality in epilepsy. Epilepsy Behav 2012;24:426–9.