Deep Brain Stimulation for Alzheimer's Disease: A Systematic Review and Meta-Analysis

Alzheimer's disease (AD) is a leading cause of dementia, characterized by progressive memory loss and cognitive decline. Medial temporal lobe (MTL) structures, including the hippocampus, parahippocampal cortices, and amygdala, undergo significant atrophy in AD, correlating with amyloid-beta and tau protein accumulation and impaired functional MRI activity during memory encoding. Atrophy in the cholinergic nucleus basalis of Meynert (NBM) and fornix degradation are also associated with cognitive impairment. Currently, there is no cure for AD, and the search for effective therapies has been challenging. Considering AD's impact on neural circuits involved in cognition and memory, neuromodulation techniques like deep brain stimulation (DBS) have emerged as potential treatment options. DBS targets neuronal activity by delivering electrical stimulation to specific brain regions. This systematic review and meta-analysis aimed to evaluate the efficacy and safety of DBS in AD, focusing on three commonly targeted regions: the NBM, the fornix, and the ventral capsule/ventral striatum (VC/VS).

The existing literature on DBS for AD showed inconsistent results across studies targeting different brain regions with varied stimulation parameters. Some studies suggested potential benefits in specific cognitive domains or subgroups, while others found no significant improvement. This inconsistency and the lack of large-scale, well-controlled clinical trials motivated the present systematic review and meta-analysis to synthesize available evidence and draw more robust conclusions.

This systematic review and meta-analysis followed the 2015 PRISMA guidelines. Databases (PubMed, Scopus, Embase, Cochrane Library, Web of Science) were searched using terms related to "Alzheimer's disease" and "deep brain stimulation." Inclusion criteria included clinical trials or reports of DBS for AD patients with a sample size of at least two patients. Data extraction involved two independent authors, with discrepancies resolved by a third senior investigator. The primary outcome measure was the effect of DBS on cognitive function, assessed using the Mini-Mental State Exam (MMSE) and/or subtests of the Alzheimer's Disease Assessment Scale-Cognitive (ADAS-cog). Risk of bias was assessed using the Cochrane Collaboration tool. Meta-analysis was performed using a fixed-effects model due to low heterogeneity (I² < 25%). Publication bias was assessed using funnel plots, trim and fill analyses, Begg and Mazumdar rank correlation, and Egger's regression intercept. Statistical significance was set at p < 0.05. Data from multiple outcomes reported by a single study were combined while avoiding bias.

The initial literature search yielded 524 publications, with five studies and six comparisons ultimately included in the meta-analysis, encompassing 66 participants. Study designs varied (pre-post designs, phase I and II RCTs, case-control). DBS targets included the NBM, fornix (± hypothalamus), and VC/VS. The meta-analysis revealed no significant overall effect of DBS on cognitive function (SMD 0.116, 95% CI −0.236 to 0.469, p = 0.518). Subgroup analyses showed non-significant effects for fornix-DBS (SMD 0.145, 95% CI −0.246 to 0.537, p = 0.467) and NBM-DBS (SMD −0.431, 95% CI −1.318 to 0.465, p = 0.341). Adverse events included neurological events (altered mental status, seizures, agitation) and non-neurological events (pain, headache). Funnel plot analysis suggested a lack of publication bias. However, the Cochrane Collaboration's tool revealed high risk of bias in many included studies due to lack of randomization and blinding. The VC/VS study showed a trend toward slower cognitive decline.

This meta-analysis did not demonstrate a significant benefit of DBS on cognitive function in AD patients. The heterogeneity in stimulation targets and parameters across studies likely contributed to these findings. The inconsistent results across individual studies targeting the fornix and NBM highlight the need for further research. The promising results from some studies, especially those targeting the VC/VS region with low-frequency stimulation, warrant larger scale RCTs. Several factors—disease stage, stimulation duration, current intensity, frequency—may influence DBS outcomes, as suggested by preclinical evidence. The mechanisms of DBS in AD may involve modulation of cholinergic neurotransmission, hippocampal theta rhythm, neurotrophic factor production, neural networks, amyloid-beta and tau levels, and neuroinflammation. The limitations of the included studies, particularly the lack of randomization and blinding, warrant caution in interpreting the results. More robust evidence from large, well-designed clinical trials is crucial to clarify the role of DBS in AD treatment.

This meta-analysis demonstrates the lack of conclusive evidence supporting the efficacy of DBS for improving cognitive function in AD. The heterogeneity in study designs, targets, and stimulation parameters, coupled with the small sample sizes and high risk of bias in many included studies, necessitates further investigation through large-scale, randomized, double-blinded, sham-controlled trials.

This meta-analysis has several limitations. The low number of included studies and small sample sizes limit the statistical power. The heterogeneity of study designs, DBS targets, and stimulation parameters makes it difficult to draw definitive conclusions. The high risk of bias in many of the included studies, due to the lack of randomization and blinding, may have influenced the results. The absence of sufficient data on specific stimulation parameters, including pulse width and amplitude, also limits the interpretation of findings. Future studies should address these limitations to provide more robust evidence.