Abstract
Background:
Antipsychotics are often used to treat neuropsychiatric symptoms in dementia, but the evidence for effect is limited. Antipsychotics have been associated with increased risk of adverse events and mortality in patients with dementia, leading to safety regulations worldwide.
Objective:
To investigate time trends in use of antipsychotics and other psychotropic drugs in dementia care.
Methods:
The study included longitudinal data on all Danish residents ≥65 years. The study population was defined on January 1 of each year from 2000–2012. Data included prescriptions, discharge diagnoses, and somatic and psychiatric comorbidities. Multivariate time trend analyses of psychotropic drug use in patients with dementia within 4-year age bands were performed.
Results:
Overall, among patients with dementia the prevalence of antipsychotic drug use decreased from 31.3% in 2000 to 20.4% in 2012. The decreasing use of antipsychotics was accompanied by decreasing use of anxiolytics and hypnotics/sedatives, but an increase in the use of antidepressants from 43.3% in 2000 to 53.8% in 2012. These changes were significant across almost all age groups. Treatment intensity among patients using antipsychotics increased as the annual median number of defined daily doses (DDD) increased from 33.3 to 42.0 DDD.
Conclusions:
The changing patterns of psychotropic drug use may be caused by warnings against use of antipsychotics. Further research is needed to explore the implications for patient safety.
INTRODUCTION
Neuropsychiatric symptoms are common in patients with dementia and include psychotic, behavioral, and affective symptoms. Up to 90% of the patients will experience one or more of these symptoms during the course of the disease [1]. Neuropsychiatric symptoms and challenging behavior are extremely distressing for patients, family, and professional caregivers and an important predictor of nursing home placement [2]. Antipsychotics are frequently used to treat behavioral symptoms in patients with dementia even though the evidence for effect is limited [3]. While treatment with antipsychotics may be necessary in selected patients, their use, and especially long-term treatment, may be associated with risks of serious adverse events and death [4, 5]. Best practice guidelines recommend non-pharmacological approaches prior to considering treatment of neuropsychiatric symptoms with antipsychotics [6, 7].
In 2005, the U.S. Food and Drug Administration (FDA) issued a warning, based on results from seventeen placebo-controlled trials, stating that the use of second-generation antipsychotics in patients with dementia was associated with a 60–70% increased mortality rate. The announcement stressed that antipsychotics were not approved for the treatment of behavioral symptoms in dementia. This was followed by a similar statement in 2008 on first-generation antipsychotics [8], and by a 2009 report from the British Department of Health stating that treating 100 patients with dementia with antipsychotics for one year would result in one death [9]. These warnings have led to safety regulations worldwide, and the Danish Health and Medicines Authority published statements in 2004 and 2008 analogous to the FDA warnings [10, 11].
Our knowledge about the use of psychotropic drugs in routine clinical practice and a possible change in prescription patterns over time is limited. Previous time trend studies in selected populations investigated use of antipsychotics before and after the warnings to evaluate the effect of a specific warning. These studies did not include, however, other confounders, such as somatic and psychiatric comorbidity, that may affect drug use in the population [12–14]. Danish registers provide a unique source for capturing complete data on drug utilization and have the advantage of eliminating selection bias. The primary aim of the present study was to investigate time trends in the use of antipsychotics in dementia care. Our hypothesis was that the prevalence of antipsychotic drug use had decreased from 2000 to 2012 in patients with dementia due to safety regulations and to the ongoing focus on health risks associated with antipsychotics. The secondary aim was to investigate time trends in the use of other psychotropic drugs and anti-dementia drugs to explore whether or not there were any compensatory changes between psychotropic drug classes in patients with dementia.
MATERIALS AND METHODS
Study design
The study was a population-based observational study using data from nationwide Danish registers.
The registers
In Denmark all permanent residents are assigned a personal identification number (CPR number) at time of birth or immigration. This makes it possible to link demographic and medical data from registers at an individual level [15]. The National Patient Registry contains data on all hospital admissions since 1977 and all contacts to outpatient clinics and emergency rooms since 1995 [16]. The Psychiatric Central Registry contains data on all admissions to psychiatric hospitals since April 1, 1969 and all outpatient contacts since 1995 [17]. The diagnoses have been registered from 1970–1993 according to the World Health Organization’s (WHO) International Classification of Diseases 8th Revision (ICD-8) and 10th Revision (ICD-10) from 1994 onwards. The Danish National Prescription Registry contains data on all dispensed prescription drugs since 1995. The drugs are registered according to the Anatomical Therapeutic Chemical (ATC) classification system, with information on the date of dispensing, package size, and strength [18].
The study was approved by the Danish Data Protection Agency (ID no.: 2007-58-0015/30-0667), Statistics Denmark, and the Danish Health and Medicines Authority (ID no.: 6-8011-907/1). Danish law did not require ethics committee approval or informed patient consent.
Study population
The population was defined for each year and included all Danish residents aged 65 or older. From January 1, 2000 to December 31, 2012, the annual population was defined at the index date January 1. Thus, an individual could be included at multiple index dates following their sixty-fifth birthday if still alive. Patients with dementia were identified as in and outpatients registered with a dementia diagnosis (Supplementary Table A1 for diagnostic codes) at discharge from a Danish hospital or at an outpatient visit and/or as individuals who had filled at least one prescription for an anti-dementia drug (ATC: N06D). Individuals who had been registered with a dementia diagnosis or filled their first prescription during the 12 months following an index date were not included in the group of patients with dementia until the next index date. Individuals diagnosed or prescribed an anti-dementia drug before age 60 were excluded from the population, because prior research has shown that the validity of dementia diagnoses in this age group is low [19]. The remaining elderly population aged 65 and older without dementia formed the reference population.
Use of psychotropic and anti-dementia drugs
Users of psychotropic drugs were defined as individuals who had filled at least one prescription for a psychotropic drug. Psychotropic drugs were classified according to ATC codes (N05A: antipsychotics; N05B: anxiolytics; N05C: hypnotics/sedatives; and N06A: antidepressants) [20]. Antipsychotics were categorized as either first-generation (levomepromazine, prochlorperazine, periciazine, haloperidol, melperone, pipamperone, droperidol, chlorprothixene, flupentixol, zuclopenthixol, pimozide, loxapine, and sulpiride) or second-generation drugs (sertindole, ziprasidone, lurasidone, clozapine, olanzapine, quetiapine, asenapine, amisulpride, risperidone, aripiprazole, and paliperidone). Lithium was analyzed separately due to its use as a mood stabilizer. Supplementary Table 2 lists all antipsychotic, anxiolytic, hypnotic/sedative, antidepressant, and anti-dementia drugs available and analyzed. The annual prevalence of patients with dementia and elderly without dementia using psychotropic drugs was calculated according to the population each year.
All drugs are assigned a defined daily dose (DDD) representing the average maintenance dose per day for a drug used for its main indication in adults, defined by WHO. The number of DDDs prescribed within a year was used to evaluate treatment duration. The mean and median DDD for users of antipsychotics among patients with dementia was calculated. As one DDD represents one day of treatment, the number of DDD was used to classify antipsychotic drugs users into two groups comprising short-term (<12 weeks) and long-term (>12 weeks) treatment. Twelve weeks of antipsychotic use was considered long-term treatment, because this duration previously was used to evaluate the efficacy and risks associated with antipsychotics [21].
Comorbidity
We used data from somatic and psychiatric hospital contacts on comorbid conditions that could potentially influence the use of psychotropic drugs. Comorbidity was evaluated at the index date every year. Registered diagnosis codes of 19 chronic somatic diseases constituted the Charlson Comorbidity Index [22]. Psychiatric comorbidity was defined as a registered diagnosis of prior psychotic disorders (one of the following diagnoses: schizophrenia, schizotypal, delusional disorders (ICD-8:295.x9, 296.89, 297.x9, 298.29-298.99, 299.04, 299.05, 299.09, 301.8; ICD-10: F20-29); manic episode and bipolar affective disorder (ICD-8:296.19, 296.39, 298.19; ICD-10: F30-F31)). The total number of different drugs used within one year was used as an indicator of somatic disease not captured by hospital admissions (ATC level 3, e.g., A10A: insulins and analogues) [20].
Statistical analysis
Time trend analysis of prevalent drug use requires independent observations. In order to achieve independency between observations, patients with dementia were divided into age groups at four-year intervals (age 65–68, 69–72, 73–76, 77–80, 81–84, 85–88, 89–92, 93–96, and 97–100). These age groups were used in the trend analysis. Drug use in each age group was compared between 2000, 2004, 2008, and 2012. Thus, an individual could not be represented at more than one point in time in the same age group. Time trend was assessed as the probability of receiving an antipsychotic, first-generation antipsychotic, second-generation antipsychotic, anxiolytic, hypnotic/sedative, and antidepressant in 2000 versus 2004, 2008, and 2012. Logistic regression analyses were performed for all age groups with use of each psychotropic drug class as outcome. We performed a crude analysis with calendar year as the independent variable and evaluated the effect of other covariates independently. Age, gender, psychiatric, and somatic comorbidity were included in the multivariate analysis (adjusted analysis) as potential confounders, because these covariates have been associated with use of antipsychotic and psychotropic drugs in patients with dementia [23, 24]. A p-value of <0.05 was considered statistically significant. During the study period, there were only 114 patients aged ≥100; therefore, this age group was not included in the trend analysis.
The data analysis was performed using SAS statistical software, version 9.3 (SAS Institute Inc., Cary, NC, USA).
RESULTS
Study population
Figure 1 shows the selection of the study population aged 65 and older in 2000 and 2012. On January 1, 2000, according to our algorithm, 19, 062 (2.4% ) patients with dementia were included, and 780, 852 individuals formed the reference population. On January 1, 2012, the number of patients with dementia had increased to 34,553 (3.5% ) and the reference population had grown to 937,537 individuals (Supplementary Figure 1 for data on 2004 and 2008). Data on nursing home placement was available from 2008 and onwards. In 2008, there were 41,234 elderly nursing home residents, 14,632 (35.5% ) of whom were patients with dementia. In 2012, out of the 45,584 elderly nursing home residents, 17,080 (37.5% ) were patients with dementia. Table 1 presents the characteristics of the study populations in 2000 and 2012. The patients with dementia were older and suffered from more comorbidity in 2012 compared to 2000.
Antipsychotic drug use
The annual prevalence of use of first and second-generation antipsychotics and other psychotropic drugs in elderly with dementia and without dementia is displayed in Fig. 2a and b, respectively. From 2000 to 2012 the annual prevalence of overall antipsychotic drug use among patients with dementia decreased from 31.3% to 20.4% . The most noticeable change occurred after 2004. In the reference population, use of antipsychotics decreased from 4.5% in 2000 to 2.8% in 2012. From 2000 to 2004, use of first-generation antipsychotics in patients with dementia decreased from 21.2% to 10.3% , while second-generation antipsychotics increased from 13.4% to 22.0% . From 2004 to 2012, use of both first and second-generation antipsychotics decreased, with first-generation antipsychotics decreasing from 10.3% to 4.5% and second-generation antipsychotics decreasing from 22.0% to 17.3% .
Results from the multivariate trend analysis are presented in Fig. 3 and detailed results can be found in the Supplementary Table 3a–f. From 2000 to 2012, antipsychotic drug use significantly decreased in all age groups, except in the oldest patients aged 97 and older. The largest change was found for patients aged 89–92, who were 50% less likely to receive antipsychotics in 2012 compared to 2000 (adjusted odds ratio (OR) [95% confidence interval (CI)]: 0.50 [0.44–0.56]). The smallest change was in patients between 65–68 years of age, who were 23% less likely to receive an antipsychotic drug in 2012 compared to 2000 (adjusted OR [95% CI]: 0.77 [0.63–0.95]).
Figure 4 shows the treatment duration of antipsychotic treatment in patients with dementia using antipsychotics (excluding lithium). There was a 46.5% increase from 2000 to 2003 (from 33.3 DDD in 2000 to 48.8 DDD in 2003), but from 2003, the median number of DDD per user stabilized at 42.0 DDD in 2012. In 2000, 18.1% of patients with dementia, who used antipsychotics, where long-term users, as they filled prescriptions for more than 84 DDDs (12 weeks). By 2012, long-term use increased to 27.3% .
Use of other psychotropic drug classes
Use of anxiolytics and hypnotics/sedatives almost halved in both groups from 2000 to 2012 (anxiolytics (2000/2012): patients with dementia 29.2% /15.7% , reference population 15.7% /7.7% ; hypnotics/sedatives (2000/2012): patients with dementia 19.2% /11.3% , reference population 18.9% /12.1% ). The decrease was significant in all age groups in patients with dementia except use of anxiolytic drugs in those ≥97 years of age. The prevalence of antidepressant drug use increased from 43.3% in 2000 to 53.8% in 2012 in patients with dementia, and from 11.1% in 2000 to 13.4% in 2012 in the reference population. Use of antidepressants increased significantly in all age groups, with the largest increase in patients with dementia aged 97–100 (adjusted OR [95% CI]: 2.88 [1.86–4.47]). In 2000, anti-dementia drugs were prescribed for 10.3% of patients with dementia and, 44.4% were in 2012.
The annual prevalence of patients with dementia not using any psychotropic drugs was also calculated. In 2000 29.8% of the patients did not receive any psychotropics, and in 2004 the prevalence was 28.3% . The prevalence of non-users increased to 33.3% in 2012.
DISCUSSION
This study investigated patterns and trends of psychotropic drug use in an entire elderly population for a 13-year period. From 2000 to 2012, the number of users of antipsychotics decreased significantly in patients with dementia to 20.4% , which is a much higher level compared to 2.8% in elderly without dementia. Even more noticeably, treatment intensity increased among patients with dementia using antipsychotics. The decreasing prevalence of use of antipsychotics in patients with dementia was accompanied by a decrease in use of anxiolytics and hypnotics, but an increase in the use of antidepressants and anti-dementia drugs.
Previous studies have reported decreasing use of antipsychotics, but aimed specifically at evaluating the impact of a particular safety warning [12–14, 26] rather than investigating their use over time in dementia care. Studies from France and Scotland that investigated time periods similar to the ones in our study found decreasing use of antipsychotics [13, 14]. However, the authors analyzed the data using a segmented regression analysis design, which did not allow adjustment for variables that might change over time [27]. Furthermore, these studies included selected populations and therefore have limited generalizability. In contrast to our study, an Italian study found that use of antipsychotics increased from 2002 to 2008 [25] in a selected population of users of anti-dementia drugs. Information about other patients diagnosed with dementia who did not receive anti-dementia treatment was missing, which limits the study’s generalizability to all patients with dementia. In a study from Germany the authors adjusted for variables that changed over time and found a small, but not significant decrease in the prevalence of antipsychotic drug use [28]. The study consisted of a series of cross-sectional studies of all patients with dementia every year, which created statistical issues as some of the observations were not independent.
Two studies investigating antipsychotic drug use at the time of dementia diagnosis have yielded varying results. A recent study from the UK found decreasing use from 1995 to 2011 of antipsychotics at the day of the first record of a dementia diagnosis and increasing use of antidepressants [29]. A Finnish study of community-dwelling persons reported that from 2005 to 2011 antipsychotic drug use within the first year of a dementia diagnosis increased [30]. However, both studies focused on treatment of prodromal symptoms of dementia and early, possibly more moderate, neuropsychiatric symptoms, whereas our study focused on treatment of neuropsychiatric symptoms at all stages of dementia.
Although we found the percentage of users of antipsychotics to decrease, the treatment intensity increased. In accordance with our study, other studies also found a stable or an increasing volume of antipsychotics prescribed for patients with dementia [13, 28]. These findings could indicate that the decreasing number of patients with dementia using antipsychotics primarily represents a decrease in sporadic or short-term users, an assumption which is supported by an increasing number of long-term users. Current Danish guidelines on antipsychotic drug use stress that antipsychotic treatment in patients with dementia should be prescribed at the lowest possible dosage for the shortest possible period of time [31]. Furthermore, it is recommended that long-term treatment (>12 weeks) is used only in cases of severe neuropsychiatric symptoms after trying non-pharmacological approaches as well as pharmacological alternatives, such as anti-dementia or antidepressants [32]. NHS guidelines also specify the importance of using short-term treatment <12 weeks [7].
Although this study was done in the Danish population, we believe that the implications of our results can be interpreted in the international context. The decreasing use of antipsychotics in Denmark and internationally is most likely the result of public debate and warnings against their use in patients with dementia. Furthermore, the downward trend may have been influenced by emerging studies concluding that the efficacy of antipsychotics is limited and that adverse events often exceed their benefits [33].
The increasing use of antidepressants in our study can be seen in the light of increasing awareness of the risks associated with antipsychotics, which might have created a need for other treatment options for patients with challenging behavior. Only few other studies have investigated potential compensatory changes in the use of other psychotropic drug classes in the light of changing prescription patterns for antipsychotics [14, 34]. In Scotland, use of antidepressants in patients with dementia doubled from 2001 to 2011 [14], but with no evidence of substitution at the time of the warnings. In the US, Kales et al. found no major compensatory changes in any individual psychotropic drug class during the warning period, but changes appeared to be spread over several drug classes [34], and the prevalence of patients not using any psychotropic drugs was constant. In our study the prevalence of non-users increased slightly from 2004 and 2012.
Overall, use of anti-dementia drugs increased from 2000 to 2012. Evidence for the efficacy of anti-dementia drugs in the treatment of neuropsychiatric symptoms is subtle [35], but Danish guidelines recommend trying treatment with anti-dementia drugs in patients with neuropsychiatric symptoms [32]. Thus, an alternative explanation may be that anti-dementia drugs have replaced antipsychotics to some degree. The significant increase in the use of antidepressants could represent a change in the way patients with dementia are treated and suggests that antipsychotics have not merely been replaced by non-pharmacological alternatives. Although the evidence for an effect of antidepressants on depression in patients with dementia is limited [36], there is some evidence that antidepressants can reduce agitation and aggressive behavior [37, 38]. One may speculate that antidepressants are more frequently used for behavioral symptoms like agitation, which raises other concerns regarding the tolerability and risk of adverse events [39]. We do not know whether changes in prescription patterns have had an effect on mortality among patients with dementia and it is also currently unknown whether there has been an effect on the quality of life of the patients.
The major strength of this study is the use of nationwide registers, which captures an entire population, thereby avoiding selection bias. A unique feature is the detailed information on drug utilization in specific age groups. We were able to include prescription data for nursing home residents, which is of great importance when investigating drug use in patients with dementia. Almost half of the patients with a registered dementia diagnosis were nursing home residents and these patients have a high prevalence of neuropsychiatric symptoms. Furthermore, the Danish registers provide data to investigate the effect of important confounders of a time trend, such as comorbidity and prior psychotic disorder. The validity of dementia diagnoses in the Danish registers has previously been shown to be high [40]. Dementia, however, is underdiagnosed, and some patients with undiagnosed or untreated dementia will have been included in our reference population. In a previous report, we found that in 2003 only 42–78% of dementia cases were registered with a diagnosis in the Danish hospital registers [41]. In addition, in the current study we included patients treated with anti-dementia drugs without a registered dementia diagnosis accounting for 11.7% of patients with dementia in 2012. Our study had further limitations, as some potential confounders were not available for the analysis, i.e., the prevalence of neuropsychiatric symptoms, level of staffing in nursing home and setting (general practice, specialized dementia or psychiatric clinic). Prescription data provides valid information about all redeemed prescriptions, but we did not have information about indications, and we do not know whether patients actually consumed the drugs they purchased. However, 85–86% of patients with dementia using antipsychotics redeemed more than one prescription, which suggests that a high number of the patients consume the prescribed medication.
The number of DDD prescribed was used to evaluate the appropriateness of antipsychotic drug use. We investigated this under the assumption of intake of one DDD per day. In the elderly and especially patients with dementia, prescribing a lower dose of antipsychotics than the standard recommendation is generally advised. Consequently, our results possibly reflect a conservative estimate of the treatment duration. We did not investigate whether users of antipsychotics received several short-term regimens or continuous prescriptions, but in the light of the risk of adverse events, the patients were exposed to a high number of doses either way.
CONCLUSIONS
This study of the entire elderly population of Denmark found a decreasing prevalence of antipsychotic drug users among patients with dementia. There was, however, evidence that many patients continued to be exposed to large amounts of the medication, which is of concern. The observed increasing use of antidepressants may also be a cause of concern, and the fact that more than 50% of all patients with dementia were prescribed antidepressants calls for further investigation. We can only speculate that antidepressants may have replaced some of the prescriptions for antipsychotics, and further research is needed to explore this. The changing patterns of drug use in dementia may be caused by the limited evidence for efficacy of antipsychotics and warnings issued by the authorities.
Footnotes
ACKNOWLEDGMENTS
The authors thank statistician Aske Astrup for his assistance with data programming and statistical analysis. The Danish Dementia Research Centre is supported by grants from the Danish Health Foundation (file no. 2007B0004) and the Danish Ministry of Health (file no. 2007-12143-112/59506 and file no. 0901110/34501). The study was supported by the Danish Medical Association (file no. 2011-3271/480853-111). All researchers were independent of the funders. All authors declare no support from any organization for the submitted work.
