Introduction

Current pharmacological treatment options for Alzheimer’s disease (AD) include the cholinesterase inhibitors (ChEIs), for mild to moderately severe AD, and memantine—an N-methyl-d-aspartate (NMDA) receptor antagonist (Kornhuber et al. 1994)—for the treatment of moderate to severe AD. At presentation, many patients are already at the moderate stage of AD (Herrmann and Gauthier 2008; Georges et al. 2008), for which ChEIs and memantine are both treatment options.

Although several controlled clinical studies with memantine have been published (Reisberg et al. 2003; Tariot et al. 2004; Peskind et al. 2006; Bakchine and Loft 2007; van Dyck et al. 2007; Porsteinsson et al. 2008), large observational post-marketing studies with significant numbers of AD outpatients and consistent use of psychometric tools are still scarce. We carried out a systematic observational study on memantine treatment for AD in Austria.

Materials and methods

Study design and patient population

This was an open-label, multicentre, prospective, 4-month observational study that investigated the efficacy and safety of memantine in outpatients with moderate to severe AD, from 88 Austrian centres (general practitioners, psychiatric and neurological practices, and outpatient hospital departments).

Patients (enroled between October 2002 and March 2004) had a diagnosis of AD (NINCDS-ADRDA criteria [McKhann et al. 1984] plus a CT/MRI scan—standard diagnostic criteria in Austria), a Mini-Mental State Examination (MMSE) score <20 (moderate to severe AD), and caregiver support. Physicians received written instructions on the NINCDS-ADRDA criteria.

Memantine was administered 20 mg/day (following 3-week titration), as recommended in the European drug approval (Memantine SPC). There were no restrictions regarding prior AD therapies; concomitant therapy was at the investigator’s discretion. The presence of clinical depression was assessed.

Assessments

MMSE score (Folstein et al. 1975) was determined at baseline and at endpoint (Month 4). The Activities of Daily Living (ADL) score (bespoke three-point rating scale comprising five items: body hygiene; eating; interest in environment; activities/hobbies; performing housework), and the Clinical Global Impression (CGI) scale (NIMH 1976; Reisberg et al. 1997) assessed non-cognitive changes at baseline, Week 4, and endpoint. Since this was an observational study, no primary efficacy variable was defined.

Safety was assessed through the evaluation of adverse events (AEs), serious AEs (SAEs) and the CGI ‘side effects’ scale.

Primary data analyses were descriptive (no alpha correction applicable, α = 0.05), with no hypothesis testing based on this observational study. Exploratory statistical comparisons were conducted between baseline and endpoint (Wilcoxon Matched Pairs signed rank test), and between the subgroups of AD treatment-naïve patients and patients switched to memantine from ChEIs (Mann–Whitney U test). Observed case (OC) analyses were applied.

According to the Austrian Drugs Act, study registration, patients’ written informed consent, and Ethics Committee centre approval were not required. However, patients could withdraw from observation at any time, on their own or their caregiver’s request, or by the investigator’s decision. SAEs or unexpected AEs were reported to the sponsor and to the Austrian Ministry of Health.

Results

Patient population characteristics

In total, 377 patients were enroled and included in the efficacy and safety analyses (Table 1).

Table 1 Patient characteristics and treatment details
Full size table

Patients were evaluated for a mean (±SD) period of 123.7 ± 41.1 days, with mean memantine doses of 20 mg/day (Week 4) and 21 mg/day (endpoint). Twenty-seven patients (7.2%) withdrew from the study, and 88.6% of patients continued memantine treatment after the observation period.

Efficacy parameters

Mean (±SD) MMSE total score improved from baseline (11.75 ± 6.38) to endpoint (13.35 ± 6.80). In patients with both baseline and endpoint data, there was a significant 1.63 ± 3.45-point improvement (exploratory analysis: p < 0.001, Wilcoxon Matched Pairs signed rank test).

Mean ADL total score improved from baseline (11.12 ± 2.42) to Week 4 (10.50 ± 2.37) and to endpoint (10.00 ± 2.49). In patients with both baseline and endpoint data, there was a significant 1.08 ± 1.86-point improvement (exploratory analysis: p < 0.001, Wilcoxon Matched Pairs signed rank test).

The subgroups of AD treatment-naïve patients and ChEI-switched patients had similar baseline MMSE and ADL measurements, with improvements in both scores at endpoint (exploratory analyses: p < 0.001, Wilcoxon Matched Pairs signed rank test). These improvements from baseline were significantly greater in treatment-naïve patients compared to switched patients (exploratory analysis: p < 0.05, Mann–Whitney U test)—Fig. 1.

Fig. 1
figure 1

Mean (±SEM) change from baseline in MMSE (a) and ADL (b) total scores after 16 weeks of memantine treatment, in AD treatment-naïve patients and patients switched from cholinesterase inhibitors (OC)

Full size image

The mean CGI severity (CGI-S) score (1 = normal; 7 = most extremely ill) improved from baseline (5.33 ± 1.01) to Week 4 (5.16 ± 1.06), and endpoint (4.96 ± 1.14). In patients with both baseline and endpoint data, there was a significant improvement of 0.38 ± 0.76 points (exploratory analysis: p < 0.001, Wilcoxon Matched Pairs signed rank test).

On the CGI global scale (CGI-C), 70.0% of patients were rated as improved (‘minimally’, 36.0%; ‘much’, 26.3%; ‘very much’, 7.7%) at endpoint; 8.5% declined, 11.7% showed no change (data were unavailable for 9.8%). Already at Week 4, 16.2% of patients were ‘much improved’ and 0.5% of patients ‘very much improved’.

Safety assessments

Forty-four patients (11.7%) reported AEs. The most frequent AEs that were possibly, probably or definitely related to memantine treatment were restlessness (n = 10), vertigo (n = 9), headache (n = 6), agitation (n = 5), nausea (n = 5), confusion (n = 4), tiredness (n = 3), and sleep disturbance (n = 3). No SAEs were reported in connection with the treatment.

On the CGI ‘side effects’ scale, side effects significantly interfered with functioning or outweighed therapeutic effect in only two patients at Week 4 and three patients at endpoint.

Discussion

While observational studies lack the power of double-blind, placebo-controlled clinical trials in determining efficacy, they can offer invaluable perspectives on the ‘real-world’ performance of approved drugs.

In controlled clinical studies of patients with moderate to severe AD (meta-analysis of six studies; n = 1,826; MMSE < 20), memantine has demonstrated significant benefits versus placebo in cognitive, functional, behavioural, and global outcomes, with a placebo-level incidence of AEs (Winblad et al. 2007). Memantine has also been shown to benefit specific cognitive items including language and functional communication (Ferris et al. 2009; Hofbauer et al. 2009; Mecocci et al. 2009), and individual activities of daily living (Doody et al. 2004). In addition, memantine-treated patients have demonstrated reduced clinical worsening (Wilkinson and Andersen 2007) and less emergence of agitation/aggression (Gauthier et al. 2005), versus placebo. Complementing these clinical results, existing observational studies found that, when added to ChEI therapy, memantine had a sustained (4-year) effect on reducing cognitive and functional decline (Atri et al. 2008) and significantly reduced the risk of nursing home admission (Lopez et al. 2009). Our study adds to these valuable findings.

At Month 4, memantine treatment produced significant improvement on the MMSE (1.63 points), ADL (1.08 points), and CGI-S (0.38 points) measures, and global improvement (CGI-C) occurred in 70% of patients. These are noteworthy improvements at this relatively advanced stage of AD given that the mean annual natural decline on the MMSE scale has been estimated at 2.3 points (Suh et al. 2004). Furthermore, treatment-related AEs were few and mild to moderate in severity, with no SAEs and a low withdrawal rate (7%).

Our results are similar to those of a German observational study (n = 1,845) in moderate to severe AD (in which memantine produced a significant 1.7-point improvement in MMSE score at 3 months (Calabrese et al. 2007)), and also extends these findings, demonstrating the onset of the treatment effect at Week 4.

Memantine produced significant benefits in mean MMSE and ADL total scores in both the treatment-naïve and ChEI-switched patient subgroups, although treatment-naïve patients showed significantly greater improvements than the switched patients at endpoint (exploratory analysis). These results suggest that memantine can produce efficacy after ChEI withdrawal, and that initiation of memantine in treatment-naïve patients is an effective alternative to ChEIs (although we cannot exclude a ‘first treatment intervention’ response). Over half of patients (56%) were AD treatment-naïve, which is consistent with reports that a high proportion of patients are in the moderate to severe stages of AD at diagnosis (Herrmann and Gauthier 2008; Georges et al. 2008).

Although benefitting from a ‘real-life’ setting, the study design has some limitations which may affect the robustness of the conclusions. For example, the study was uncontrolled (statistical analysis was exploratory), open-label, and employed subjective real-world entry criteria. Some sources also criticise the MMSE test as an outdated cognitive assessment, and objective tools for cognitive assessment may display floor effects in patients with more advanced dementia. However, the MMSE is widely used by practising physicians, and Austrian pharmaceutical law does not permit the use of any test that could be perceived as beyond clinical routine.

In conclusion, the results of this observational post-marketing study support the findings of controlled clinical trials, and reveal that in ‘real-life’ treatment of moderate to severe AD, memantine is efficacious with respect to cognitive, functional and global outcomes, and is well-tolerated. These positive results are seen in patients receiving memantine as their first AD pharmacotherapy, and in patients switched from ChEIs. Exploratory analyses suggest that the effects of memantine may be optimised by early treatment initiation.