Neuropsychiatric symptoms in patients with dementia in Norwegian nursing homes : the course of the symptoms and the effect of discontinuation of psychotropic medication

Neuropsychiatric symptoms (NPS) are prevalent in dementia, and it has been estimated that up to 90% of people with dementia will experience NPS in the course of their dementia. Sometimes NPS in dementia are poorly diagnosed, and the effect of the treatment prescribed for the NPS has been poorly documented. The aim of this thesis was to describe the prevalence, course and treatment of NPS in Norwegian nursing homes. We wanted to study whether NPS are transient or persistent. In addition, we wanted to investigate the effect of discontinuing treatment with antidepressants and antipsychotics on patients in Norwegian nursing homes with dementia and NPS. We conducted a small pilot discontinuation study, and a large double blind randomised controlled discontinuation trial (DB RCT). Previous DB RCT discontinuation studies of antipsychotics have shown that discontinuation of the medication has been beneficial for the patients, while the effect of the discontinuation of antidepressants in patients with dementia and NPS has not previously been studied in a DB RCT. To obtain good assessments of the cognitive function of the patients in the study, most of them having moderate or severe dementia, we translated the Severe Impairment Battery (SIB) into Norwegian and validated this instrument. Four studies were conducted. In the validation study of the SIB 59 patients from three nursing homes in Hedmark and Oppland were included. In the study on the prevalence and the course of NPS in nursing homes, 210 patients from seven nursing homes in Hedmark and Oppland were included. In the pilot discontinuation study, 23 patients from seven nursing homes in Hedmark and Oppland were included, while in the DB RCT discontinuation study of antidepressants 128 patients from 52 nursing homes in 14 counties of Norway were included. The Severe Impairment Battery (SIB) is a cognitive test for patients with moderate to severe dementia (minimum score 0 and maximum score 100). In the SIB validation study, three psychiatrists/doctors drafted the cognitive questionnaire into Norwegian before a psychiatrist made a final Norwegian translation from the three drafts. A colleague born in Newcastle, who has lived in Norway for several decades, translated the Norwegian version of the SIB back into English, and this version corresponded well with the original English version. The patients' cognition and degree of dementia were examined with the SIB and the Clinical Dementia Rating scale (CDR), while the patients were diagnosed with dementia according to the International Classification of Diseases (ICD-10). A nurse and a doctor used the SIB to assess the patients within seven days of each other. The average SIB score was 72.10 points. In the reliability study Cronbach's alpha was 0.97, indicating a good internal reliability of the test. Spearman's rho correlation coefficient between the two testers was 0.85 for the total SIB score, and ranged between 0.46 and 0.76 for the sub-items of the test, which indicated a good inter-rater reliability. Scoring results on the SIB were compared with the CDR scores of patients. Spearman's rho correlation coefficient between the SIB score and the CDR score was 0.55. The groups of patients with CDR 1, 2 and 3 were significantly different from each other as measured by the SIB. By using Receiver Operating Characteristic (ROC) curve analysis we found that the SIB score of 87/88 best discriminated between CDR In the study of the prevalence and the course of NPS, patients were examined at baseline (T0) and after four (T1) eight (T2), twelve (T3) and sixteen (T4) months with the Neuropsychiatric Inventory (NPI), CDR, the Mini-Mental State Examination (MMSE) and the Quality of Life in late-stage dementia (QUALID) scale. The NPI examines NPS, and we reported the prevalence and course of the NPS identified with the NPI. At baseline, the average age of the patients included was 84.9 years, 69.2% were female and the median length of stay in the nursing home was 673 days. The three most prevalent NPS were irritability, agitation/aggression and apathy (T0 and T1), irritability, agitation/aggression and disinhibition (T2 and T3) and depression, disinhibition and irritability (T4). Irritability had the highest cumulative prevalence (63.5%), followed by agitation/aggression (51.0%) and disinhibition (50.0%). In total, 91.7% of patients had at least one NPS during the 16 months period. Irritability (42.6%), disinhibition (37.8%) and depression (31.5%) had the highest cumulative incidence. The most persistent NPS were agitation/aggression, irritability and disinhibition (T0-T1) and (T1-T2), disinhibition, apathy and irritability (T2-T3) and hallucinations, depression and anxiety (T3-T4). The NPS with highest resolution rate were euphoria, appetite and eating disorders, and sleep and night-time behaviour disorders (T0-T1) and (T1-T2), appetite and eating disorders, hallucinations and delusions (T2-T3) and appetite and eating disorders, euphoria and apathy (T3-T4). The conclusion of this study is that almost all patients included in this study have one or more NPS in the course of sixteen months, but individual symptoms fluctuate, which should affect the treatment which patients are given. In the small pilot study where we investigated the effect of the discontinuation of antidepressants and antipsychotics, 23 patients with dementia, but without a depressive disorder, were included. Twelve patients used antipsychotics of different types and 11 patients used selective serotonin reuptake inhibitor (SSRI) antidepressants. Patients were examined at baseline with the CDR, the NPI, the Cornell Scale for Depression in Dementia (CSDD), a sub-scale of the Unified Parkinson's Disease Rating Scale (UPDRS), the SIB, the Lawton and Brody's Physical Self-Maintenance Scale (PSMS) and the Quality of Life-Alzheimer's Disease (QoL-AD) scale. At three, six and 12 weeks, patients were examined with the NPI and the UPDRS, and after 24 weeks the same assessment scales as at baseline were used. At inclusion the average age was 84.1 years and 91.3% were women. At three and six weeks, we found a small increase in the NPI and the UPDRS scores of patients in both groups, but at 12 and 24 weeks the scores on both scales were back to baseline levels. The SIB scores (cognition) at 24 weeks showed a slight decrease in the group who discontinued antidepressant medication and a small increase in the group who discontinued antipsychotics. The CSDD scores were unchanged in the antipsychotic group, but showed a small decrease in the antidepressant group. None of the results were statistically significant, but suggested that discontinuation of antidepressants and antipsychotics were safe in the patients with dementia and NPS, and could even be beneficial for the patients. In the randomised double-blind RCT discontinuation study of antidepressants, the 128 patients were assessed with the same assessment tools as in the pilot study and the assessments were done at baseline and after four, seven, 13 and 25 weeks. The 128 patients used escitalopram, citalopram, sertraline or paroxetine at inclusion. In half of the patients the antidepressive medication was discontinued, while in the other half of the patients the medication was continued. The study was double blind and placebocontrolled, meaning that in patients who discontinued medication the antidepressant was substituted by placebo tablets or capsules identical in appearance to the study medication in those who continued with medication. Neither the patients, the relatives, employees at the nursing home or the study management knew which patients were in the discontinuation group and which were in the active medication group, as randomisation and distribution of the study medication was made by the hospital pharmacy at Innlandet Hospital Trust, Gjøvik. Randomisation was by a computer-generated 1:1 allocation sequence. Sixty-three patients were allocated to the antidepressants discontinuation group (ADG), out of which 27 patients (42.9%) had to discontinue the study within 25 weeks. Sixty-five patients were allocated to the antidepressants continuation group (ACG), out of which 18 patients (27.7%) had to discontinue the study within 25 weeks. At baseline, the groups were comparable in terms of age, gender, NPI score, CSDD score and all the other measured variables. At 25 weeks, the ADG group had an increase of 2.53 points on the CSDD from 5.03 points while the ACG group had a decrease of -0.43 point from 5.89 points. The difference between the groups was statistically significant. A post hoc analysis showed that the difference in the total CSDD score between the groups were statistically significant different as early as at week 7 (visit 3). Measured with the NPI the ADG group had an increase of 5.93 points from 17.79 points while the ACG group had a decrease of - 1.39 points from 17.63 points. When analysing the results with ANCOVA, and correcting for the baseline values of the CSDD, the difference between groups was still statistically significant. We analysed the mood sub-scores of the CSDD, based on a Norwegian factor analysis (Barca et al., 2008), and the affective subscales of the NPI (NPI-depression + NPI Anxiety) to examine the mood symptoms of the patients. This analysis also showed statistically significant differences between the groups. For the other assessment tools – the CDR, the UPDRS, the QoL-AD and the PSMS – there were no statistically significant changes between baseline and 25 weeks. We concluded that discontinuation of antidepressants in patients who have dementia and NPS, but no depressive disorder, led to an increase in depressive symptoms