Criteria for participation in the study included written informed consent as well as caregiver consent, a diagnosis of probable AD by means of NINCDS-ADRDA criteria (McKhann et al. 1984), a Mini Mental Status Examination score <25, and receipt of a 5–10 mg dose of donepezil for at least 2 months. All patients continued to receive donepezil during the study. To be enrolled in the study patients had to be 50–80 years of age and not able to bear children. Patients with severe and unstable somatic diseases and patients with present or known history of alcohol, drug or medication abuse were excluded. Patients taking drugs for coexistent diseases were included except those taking psychotropic drugs, “nootropics” or health food supplements. The trial has been approved by the appropriate ethics committee and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All persons gave their informed consent prior to their inclusion in the study.
The study had a monocenter, prospective, double-blind, placebo-controlled, parallel-group randomized design. Sixty-eight patients were recruited. Patients were randomly allocated either to placebo or to verum (Cu orotate). The oral dosage of Cu-(II)-orotate-dihydrate was 51.62 mg (corresponding to 8 mg Cu) once daily. To maintain blinding, capsules containing Cu orotate and placebo had identical shape and color. The duration of the study was 12 months. The study was registered at www.clinicaltrials.gov with the identification number NCT00608946.
Screening procedures consisted of a medical history, physical examination and psychometric tests. Blood was collected to determine the level of Cu and zinc by using atomic absorption spectroscopy (AAS). Furthermore, blood count included liver enzymes and anti-mitochondrial antibodies (for exclusion of cholestatic liver diseases). Atomic absorption spectroscopy: Blood samples were collected into metal-free tubes that contained lithium heparin as an anti-coagulant. Concentrations of Cu were measured in lithium-heparin plasma samples, utilizing flame AAS (Perkin Elmer, AAnalyst 800). Samples were diluted with deionized water and the analysis was performed against standards prepared in glycerol to approximate the viscosity characteristics of the diluted samples. Standard atomic absorption conditions were utilized for Cu (air acetylene, wave length 324.8 nm).
The primary efficacy variables were the changes from the baseline score on the cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-cog) and on the MMSE. Efficacy evaluations were performed at baseline and at month 3, 6, 9 and 12.
Therapeutic drug monitoring
Plasma Cu levels were measured at every visit (at baseline and at month 3, 6, 9 and 12).
Standard adverse event reporting was conducted. Before baseline, the Cu content in drinking water at the patients’ home was analyzed under standard conditions; concentration of Cu in drinking water must not exceed 2 mg/l. For safety reasons, blood analyses (including especially plasma Cu levels and liver enzymes) were performed at week 1, 2, 3 and 4 after baseline and at month 3, 6, 9 and 12.
Evidence of efficacy was indicated by a significant difference in change from baseline between treatment arms. For statistical analyses SPSS 14 was used. All tests were two-tailed. Demographic variables, data on the course of the disease and clinical data at baseline of the patients that completed the examinations were compared between the verum and the placebo group with one-way analysis of variance (ANOVA). Chi-square test on independence was used to analyze, if the distribution of sex was different between the two groups. Dependent variables were the ADAS-cog total score and the MMSE score at baseline (t
0) and at four following measurements with intervals of 3 months each (t
12). Furthermore, plasma Cu levels at the beginning (t
0) and at the end (t
12) of the study were compared. Independent between-subject factor was treatment group (placebo, verum), within-subject factor was time of measurement. For ADAS-cog, MMSE and plasma copper the differences between the values at the end (t
12) and at the beginning of the study (t
0) were computed as measures for the changes over time. Pearson’s product moment correlations between these differences and intervening variables age, disease duration, age at onset of the disease and duration of education were calculated, intending to analyze, if these variables show a significant influence on the changes of the dependent variables. ANOVA with between-subject factors treatment group and gender was use to analyze, if the changes were gender-related. For the dependent variables Kolmogorov–Smirnov tests were used to test, if there were significant deviations from the normality assumption. The results of these preliminary statistics regulated the main analysis. The basic hypotheses were that the dependent variables ADAS-cog and MMSE change continuously over time and that these changes are different between the placebo and the verum group. The third hypothesis was that plasma copper level increases over time in the verum, but not in the placebo group. To analyze these hypotheses, the general linear model procedure was used to perform multivariate analyses of variance (MANOVA) with repeated measures design. Within-subject factor was time of measurement between-subject factor was treatment group. Since there was no significant influence of the intervening variables sex, age, disease duration, age at onset of the disease and duration of education on the differences of the dependent variables between the values at the end and at the beginning of the study, no additional covariates or factors were entered to the model in the main analysis. In the repeated measures analyses MANOVA performed an orthonormal transformation of the dependent variables. The Mauchly test of sphericity was used to test the assumption that the covariance matrix of the transformed orthonormal variables was constant on the diagonal and zero off the diagonal. This was the case, if the values of the dependent variables show the same error and interaction effects at each measure time. If the sphericity assumption was violated, the Greenhouse–Geisser epsilon correction was applied.