Psychotropic and anti-dementia treatment in elderly persons with clinical signs of dementia with Lewy bodies: a cross-sectional study in 40 nursing homes in Sweden

The study material consists of data from the dementia NH residents in Malmö, the third largest Swedish city with over 320,000 inhabitants. To gather accurate information on the actual number of dementia NHs, for the period 2012–2013, we contacted the head account manager for dementia NHs in the city of Malmö. Thereafter, from January 2012 to March 2013, we invited all 40 dementia NHs in Malmö to participate in the study. All residents aged ≥65 years had the same electronic hospital medical record system and updated medical lists during the study period.

The NHs in Malmö each housed an average of 15 elderly residents (range 4–45) with a staff consisting of a resident physician (specialist from primary care), 1 or 2 head nurses (with 3 years of university) and a varying number [4‐8] of assistant nurses (with 2 years of nursing care school). The elderly residents either already had a dementia diagnosis or were undergoing a dementia investigation procedure. All resident physicians are requested to follow the guidelines for dementia investigation from the Swedish NBHW [30]. The minimal investigation, the so-called “Basal dementia investigation”, consists of cognitive testing (Mini–Mental State Examination, cube and clock drawing tests), relevant laboratory tests, cone beam computed tomography (CBCT) and/or magnetic resonance (MR) imagining (evaluated by a radiologist) and structural anamnestic interview by the doctor (patient and family members).

The elderly person becomes an NH resident after a decision made by a care organizer from the local authority (communal/municipal) following care planning with the elderly person, the family members and a doctor or head nurse who had consulted with the doctor.

The inclusion criterion was all residents living at NHs ≥65 years. The exclusion criterion was residents in a palliative state of treatment. The main clinical outcome measure was a questionnaire that was specially designed to cover clinical signs of DLB according to the consensus criteria from the third report of the DLB Consortium [7] but also compatible with the diagnostic form of DLB Diagnostic Symptoms of the Lewy Body Dementia Association [31]. We decided to choose observable descriptions of possible DLB signs. Prior to entering the study, each NH separately received a specially designed presentation seminar given by a resident physician (IZ) to all nursing staff. For this study, there was also the opportunity of individual medical education, clarification and answering questions from the nursing staff (head nurses and assistant nurses). There were also possibilities for the resident physician to demonstrate to the nursing staff how the clinical features from the questionnaire could be expressed in practice in the elderly (with their verbal consent) living in NHs. The nursing staff were instructed to register the presence/absence of seven actual clinical features at the time of scoring: 1) Parkinson’s disease (PD) or parkinsonism; 2) rigidity, signs of stiffness; 3) tremor and/or a rigid posture and walk; 4) a weak voice; 4) balance problems, dizziness, faints, falling easily and/or more frequently; 5) visual hallucinations; 6) excessive daytime sleepiness, variations in attention and wakefulness; and 7) sleep disorder, acting out dreams during sleep (sometimes violently), shouting out at night, nightmares. A more detailed description of the questionnaires is presented in an earlier publication [29].

During the same period, the head nurse collected actual medical lists from the Swedish National Medication Dispensing System (NMDS) together with the questionnaire. From the computerized hospital medical records, we collected data on dementia diagnoses (yes/no and dementia type). From the medical lists, we categorized the psychotropic medicines according to the Anatomical Therapeutic Chemical (ATC) Classification System: antipsychotics (N05A), anxiolytics (N05B), hypnotics/sedatives (N05C), anti-depressants (N06A) and anti-dementia medicine (N06D) [32].

For the statistical analysis, we used the IBM Statistical Package for the Social Sciences (SPSS) for Windows (version 22.0; IBM Corporation, Armonk, NY, USA). For comparisons of two independent groups, we used the Student’s t test for normally distributed data and the Mann–Whitney U test for non-parametric data. NH residents were divided into dichotomous groups: dementia (yes/no), DLB signs (0–1 DLB and 2–4 DLB) and age (≤85 and ≥86 years) (Tables 1, 2 and 3). Categorical data were investigated using the chi-squared test. The significance level was set to P < 0.05. A Bonferroni correction for multiple testing of the groups was applied by setting the P-value at < 0.01.

In this study, 40 NHs with 650 residents were invited to participate. Of these, 30 residents declined participation and 10 had missing data. Thus, the final study population consisted of 610 elderly residents with collected medical lists for 576 (94%), hospital medical records for 594 (97%) and completed questionnaires for all residents. The mean age was 86 (± 6.9) years and 76% were women; 74% of residents had a dementia diagnosis (Table 1).

The number of different DLB signs, from the questionnaires, were used to create two groups: residents with 0–1 DLB signs and those with 2–4 DLB signs. In an earlier study of the same population, we reported that 84% had 0–1 and 16% had 2–4 main clinical features of DLB [29]. The residents with 2–4 DLB signs had the following dementia diagnoses: AD in 19%, DLB/Parkinson’s disease with dementia (PDD) in 17%, dementia not otherwise specified (NOS) in 16%, vascular dementia (VaD) in 13%, AD mixed dementia (AD-MIX) in 12% and no diagnosis in 23%. Finally, 22 of the residents were already diagnosed with DLB/PDD, 20 (91%) of these were classified within the 2–4 DLB signs group (Table 1).

We found the usage of any psychotropic medication in 86% of residents. All four psychotropics (N05A, N05B, N05C and N06A) were found in 6%, three in 17%, two in 31%, and one in 32% of residents. We found the usage of anxiolytics in 58%, hypnotics/sedatives in 41%, anti-depressants in 50% (Table 2). In addition, anti-dementia medication in 33% of residents (Table 3).

Usage of antipsychotics was found in 22% of the study population, most prevalently risperidone (10%) followed by haloperidol (7%) and other atypical antipsychotics (clozapine, olanzapine and quetiapine) (2%) (Table 2). There was no difference in antipsychotic treatment according to sex or dementia diagnosis (yes/no). However, the age groups showed differences where younger residents (≤85 years) had significantly more frequent usage of antipsychotics compared with those ≥86 years (Tables 2 and 3). Furthermore, the usage of antipsychotics increased with the increasing number of main DLB signs from 25% to 43% (Fig. 1). Comparing all residents, those in the 2–4 DLB group had the highest usage of any antipsychotics (32%) and the highest usage of haloperidol (14%) (Table 2).

Of the participants with a dementia diagnosis, 45% had treatment with some anti-dementia medication. None of the residents without a dementia diagnosis had anti-dementia medication (Table 3). Younger residents (aged ≤85 years) had significantly higher usage of some anti-dementia medication (43%) compared with those aged ≥86 years (26%) (Table 3).

Comparing the 0–1 DLB and 2–4 DLB groups, there were no differences in treatment with any cholinesterase inhibitors (ChEI). However, higher usage of memantine was seen in the 2–4 DLB group (Table 3). In the group with 2–4 DLB signs, 37% were undergoing anti-dementia treatment, while 62% of those already with a diagnosis for DLB/PDD were undergoing anti-dementia treatment.

A number of studies have emphasized the importance of identifying DLB/PDD patients at an early stage, which in turn may reduce the risk of an incorrect diagnosis, unsuitable medication, worsening of the quality of life, early NH admission and increased risk of shorter survival [33‐37]. By dividing the study population into two groups (0–1 and 2–4 DLB signs), we could extract the high-risk residents for DLB/PDD diagnosis together with their clinical features, medical usage and actual diagnoses. The prevalence of NH residents with 2–4 DLB signs in this study was 16%. Although not specifically addressed here, the observational data from clinical nurses and recorded clinical signs at the time of the study suggest that residents with 2–4 DLB signs could have undiagnosed DLB/PDD. Furthermore, over 90% of the residents already with a diagnosis for DLB/PDD from hospital medical records were in the group with 2–4 DLB signs, which indicates that the study questionnaire was efficient in identifying residents with possible DLB/PDD (Table 1).

By using the Swedish NMDS and the hospital medical records, together with the actual clinical features reported by nursing personnel who know the residents best, we are able to present accurate medical usage for this large study population and contribute to the quality of the analysed data. Furthermore, the dementia diagnoses were investigated by resident physicians who must follow basal guidelines for dementia investigation from the Swedish NBHW [30].

Nevertheless, the study had a number of limitations. It was not possible to investigate the severity of dementia of residents due to the lack of this information. Since severe dementia according to ICD-10 is defined as a need of permanent support and caregiving from others, all residents of this study were classified as having severe dementia. Another limitation was the lack of information about why 170 elderly residents were “undiagnosed” NH dementia residents. We assume that these undiagnosed residents either were under basal dementia investigation or did not have a dementia diagnosis that was registered at the time of our data collection.

The usage of psychotropics is still high and according to several Swedish, Danish and Norwegian studies varies greatly (39%–71%) due to whether the elderly are institutionalized or not and managed by primary care or specialist clinics [25, 27, 28]. In our study, we show that antipsychotics, anxiolytics, hypnotic/sedatives and anti-depressants were used in the NHs to the same degree as the above mentioned with no difference by sex or dementia diagnosis (Table 2).

By dichotomizing the residents into two age groups (≤85 and ≥86 years), we showed that those ≤85 years were more frequently treated with antipsychotics (both haloperidol and risperidone), which may be appropriate according to national recommendations for treatment.