The impact of multidomain interventions on cognitive and physical function in older adults with subjective cognitive decline: a meta‑analysis and systematic review

Introduction

With the global population aging, age-related cognitive decline has become a major public health concern. It is predicted that 152.8 million individuals will experience dementia by 2050 (GBD 2019 Dementia Forecasting Collaborators, 2022), leading to a considerable impact on the socio-economic system (Alzheimer’s Association, 2023). Subjective cognitive decline (SCD) is defined as self-reported cognitive complaints in the absence of objectively measurable cognitive deficits on neuropsychological assessments (Jessen et al., 2014). SCD is commonly regarded as the earliest precursor stage of Alzheimer’s disease, and as an intermediate stage between normal aging and mild cognitive impairment (MCI). A meta-analysis of 28 cohort studies found that individuals with SCD had a two-fold relative risk of 2.07 (95% CI [1.77–2.44]) for dementia conversion compared to those with healthy cognitive aging. The annual conversion rates were 2.33% for SCD and 6.67% for MCI (Mitchell et al., 2014). Those findings highlight the importance of early interventions, with SCD offering a promising window to delay further cognitive decline.

Currently, the efficacy of pharmacological interventions for cognitive decline remains unproven, and they commonly have side effects (Livingston et al., 2020). In contrast, non-pharmacological interventions have demonstrated promising effects without adverse effects for older adults with SCD (Smart et al., 2017). Among these, nutritional, exercise, and multidomain interventions are three effective non-pharmacological approaches to improve cognitive function. Nutrition is a key modifiable factor in preserving brain health during aging (Flanagan et al., 2020; Melzer et al., 2021; Morris, 2012). The cognitive benefits of nutritional interventions have been examined through various nutrients and dietary patterns, including folic acid (Gil Martínez, Avedillo Salas & Santander Ballestín, 2022), vitamin D (Jiang et al., 2023), and omega-3-polyunsaturated fatty acid (Cherbuin, Anstey & Baune, 2017), the Mediterranean diet (Valls-Pedret et al., 2015), the Dietary Approaches to Stop Hypertension diet (DASH) (Smith et al., 2010), and ketogenic diet (Krikorian et al., 2012). Nutritional interventions improve cognitive function primarily through the supplementation of specific nutrients. Diets rich in antioxidants, healthy fats, and anti-inflammatory compounds—including the Mediterranean and DASH diets—have been shown to reduce oxidative stress and inflammation, improve vascular health, and protect against age-related neurodegeneration (Morris et al., 2015; Scarmeas et al., 2007).

Physical exercise is another well-established modifiable factor in preserving cognitive health. Aerobic (Kamijo et al., 2009; Young et al., 2015), resistance (Coelho-Junior et al., 2022; Liu-Ambrose et al., 2010), multicomponent exercises (Li et al., 2021), and mind-body exercises (Wu et al., 2019; Xia et al., 2019) are common strategies in preserving cognitive function and preventing cognitive decline in older adults. Physical activity may indirectly influence cognitive function by improving overall health and enhancing neuroplasticity (Bherer, 2015; Bruderer-Hofstetter et al., 2018). Similarly, cognitive interventions, based on the theories of neuroplasticity, have proven to be effective strategies for mitigating cognitive decline (Liu et al., 2021a; Marlats et al., 2020). Some systematic reviews have demonstrated the positive effects of multidomain interventions in enhancing cognitive outcomes (Gavelin et al., 2020; Sood et al., 2019). Cognitive interventions provide a targeted approach to improve brain function by directly stimulating and challenging cognitive processes. Cognitive interventions may directly impact cognitive function by enhancing cognitive plasticity, cognitive reserve, and neural connectivity (Cheng, 2016; Li et al., 2017; van Balkom et al., 2020).

Multimodal intervention strategies have gained increasing attention for their potential synergistic effects. Multidomain interventions are defined as interventions that incorporate at least two different domains, such as nutritional, exercise, cognitive, or psychosocial interventions (Bevilacqua et al., 2022). Currently, the more effective multidomain intervention patterns are those that include nutritional, exercise, and cognitive interventions (Castro et al., 2023), as demonstrated by large-scale trials such as the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) (Ngandu et al., 2015) and Multidomain Alzheimer Preventive Trial (Andrieu et al., 2017). The Japan-Multimodal Intervention Trial for Prevention of Dementia (J-MINT) is also a part of the global FINGERS network. The study found that, after 18 months, the intervention group showed a significant improvement in cognitive composite scores (Oki et al., 2024). In addition, MIND-AD trials targeting the early stages of Alzheimer’s patients demonstrated that interventions combining lifestyle modification and drug supplementation reduced the rate of cognitive decline after 6 months (Thunborg et al., 2024). Furthermore, a meta-analysis suggested that multidomain interventions combined with physical exercise might benefit overall cognitive function in older adults (Reparaz-Escudero et al., 2024). Conversely, the AgeWell.de trial adjusted the FINGER intervention model to test a multimodal intervention (including optimized nutrition, medication, and physical, social, and cognitive activity), but no effect on global cognitive performance was detected (Zülke et al., 2024). Similarly, another study found no evidence supporting the effectiveness of multidomain interventions for preventing dementia (Hafdi, Hoevenaar-Blom & Richard, 2021). Overall, the effects of multidomain interventions on cognitive function in older adults remain controversial (Hafdi, Hoevenaar-Blom & Richard, 2021; Oki et al., 2024; Thunborg et al., 2024; Zülke et al., 2024), and comparisons with single-domain interventions are still limited (Dedeyne et al., 2017; Fessel et al., 2017).

Although some meta-analyses and reviews have reported positive effects of multidomain interventions on cognitive and physical functions (Ahn et al., 2022; García-Llorente et al., 2024; Liao, Shen & Li, 2023; Mohanty & Kumar, 2022), the varied components of these included interventions may introduce significant heterogeneity. In addition, these studies included designs other than randomized controlled trials (RCTs) (Fessel et al., 2017; Mohanty & Kumar, 2022) and lacked rigorous inclusion criteria for study populations (Liu et al., 2021b; Noach et al., 2023). To our knowledge, only one review has examined the effects of multidomain interventions on individuals with SCD, targeting those over 45 years old and including RCTs, cohort studies, and reviews (Mohanty & Kumar, 2022). However, no meta-analysis with robust included criteria exists for older adults with SCD. Therefore, this study defines multidomain interventions as a combination of nutrition, physical exercise, and/or cognitive training, and includes only RCTs to examine their effects on cognitive and physical function in older adults with SCD.

This is the first meta-analysis with a robust design to examine the effectiveness of the multidomain interventions on cognitive and physical functions among older adults with SCD. This study aimed to evaluate the effects of multidomain interventions compared to the control and nutritional interventions.

Materials and Methods

This meta-analysis complied with the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) guidelines (Liberati et al., 2009) and PRISMA extension for meta-analysis (Hutton et al., 2015). The present study has been registered on the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY, registration number: INPLASY202460020).

Results

Study selection

As shown in Fig. 1, the initial search identified 5,402 records. After removing 1,080 duplicates, the two researchers screened the remaining 4,322 documents by reading the titles and abstracts. A total of 133 articles were then selected for full-text review, resulting in the exclusion of 128 articles that did not meet the inclusion criteria. The specific reasons for exclusion are provided in Fig. 1. Ultimately, five studies met the inclusion criteria. In addition, one eligible article was identified by searching the references of the included studies. Thus, a total of six articles were included in the study.

Risk of bias of included studies

As depicted in Fig. 2, the bias assessment follows the guidelines outlined by the Cochrane Collaboration (Higgins & Green, 2008). The graph utilizes green, yellow, and red colors to represent low risk, unclear risk, and high risk, respectively. Among the six included studies, five studies (Andrieu et al., 2017; Barreto et al., 2021; Chatterjee et al., 2022; Liu et al., 2023; Macpherson et al., 2022) reported detailed methods for generating random sequences, and three (Andrieu et al., 2017; Barreto et al., 2021; Chatterjee et al., 2022) provided the methods for allocation concealment. The blinding of participants and personnel in the four studies was identified as low risk (Andrieu et al., 2017; Barreto et al., 2021; Chatterjee et al., 2022; Macpherson et al., 2022). Five studies (Andrieu et al., 2017; Barreto et al., 2021; Blumenthal et al., 2020; Liu et al., 2023; Macpherson et al., 2022) described the use of blinding of outcome assessment. Notably, three studies (Barreto et al., 2021; Liu et al., 2023; Macpherson et al., 2022) were at high risk of selective reporting. Lastly, all six studies (Andrieu et al., 2017; Barreto et al., 2021; Blumenthal et al., 2020; Chatterjee et al., 2022; Liu et al., 2023; Macpherson et al., 2022) reported no risk of incomplete outcome data and other bias.

Effects of multidomain interventions vs the control group

Global cognition

Four studies (Andrieu et al., 2017; Barreto et al., 2021; Liu et al., 2023; Macpherson et al., 2022) investigated the effects of multidomain interventions on global cognition. Global cognition was assessed using the Mini-Mental State Examination (MMSE) or Montreal Cognitive Assessment (MoCA). This analysis used the SMD due to the different assessment tools used across studies. The pooled results indicated that multidomain interventions did not significantly improve global cognition with a small effect size (SMD = 0.08, 95% CI [−0.03 to 0.20], p = 0.15, Fig. 3).

Executive function

Three studies (Chatterjee et al., 2022; Liu et al., 2023; Macpherson et al., 2022) evaluated the effectiveness of multidomain interventions vs the control group. As shown in Fig. 4, multidomain interventions significantly improved executive function, with a moderate effect size (SMD = −0.24, 95% CI [−0.45 to −0.03], p = 0.02).

Memory

Figure 5 demonstrates four studies (Andrieu et al., 2017; Barreto et al., 2021; Chatterjee et al., 2022; Liu et al., 2023) that examined the memory gains of multidomain interventions. The pooled analysis indicated that multidomain interventions greatly enhanced memory with a moderate effect size (SMD = 0.45, 95% CI [0.03–0.88], p = 0.04).

Physical function

Two studies (Andrieu et al., 2017; Barreto et al., 2021) assessed the efficacy of the multidomain interventions on physical function using the Short Physical Performance Battery. Pooled results were not statistically significant with a small effect size (MD = 0.02, 95% CI [−0.25 to 0.29], p = 0.88, Fig. 6).

Effects of multidomain interventions vs nutritional interventions

For comparisons with single domain intervention, due to the limited number of studies, the meta-analysis could not be performed on separate exercise interventions and cognitive interventions. Thus, we only compared the effects of multidomain interventions with those of single nutritional interventions.

Executive function

Two studies (Andrieu et al., 2017; Blumenthal et al., 2020) evaluated the effectiveness of multidomain interventions vs nutritional interventions. As shown in Fig. 7, multidomain interventions produced a significantly enhanced executive function, with a moderate effect size (SMD = −0.31, 95% CI [−0.59 to −0.04], p = 0.03).

Memory

Figure 8 shows that two studies (Andrieu et al., 2017; Blumenthal et al., 2020) examined the effects of multidomain interventions on memory. The pooled results reported that multidomain interventions significantly improved memory with a moderate effect size (MD = −4.07, 95% CI [−6.71 to −1.31], p = 0.004).

Discussion

To our knowledge, this is the first study to examine the efficacy of multidomain interventions on both the cognitive and physical functions of older adults with SCD. The results indicated that multidomain interventions significantly improved executive functions and memory, but did not impact global cognition and physical performance when compared to the control group. When compared to nutritional interventions, our study demonstrated that multidomain interventions significantly enhanced executive function and memory.

This study demonstrated that multidomain interventions have a moderate impact on executive function and memory compared to the control group, which may hold clinical relevance for older adults with SCD. These findings align with the review by Ahn et al. (2022), which included seventeen RCTs examining the effects of multidomain interventions on executive function and episodic memory in older adults without dementia (Ahn et al., 2022). Notably, executive function and memory have complex subcomponents (Wang & Dong, 2018), and different reviews may use various tools to assess these components, which would increase the heterogeneity of pooled results. Regarding global cognition, our study found that multidomain interventions could not significantly improve global cognition, contrary to previous results (Ahn et al., 2022; Liao, Shen & Li, 2023). A possible explanation is that the intervention period in this study was relatively short, which may have limited the manifestation of intervention effects. In addition, the limited number of included studies may also influence the pooled results.

In comparison to single-domain interventions, we only compared the effects of multidomain interventions with single nutritional interventions due to limited included studies. Our results indicated that multidomain interventions effectively enhanced executive function and memory with a moderate effect size, suggesting a possible synergistic moderate effect. A review reported that multidomain interventions are more promising than single-domain interventions, which supports our results (Fessel et al., 2017). However, our results should be interpreted with caution, as they are based on only two studies (Andrieu et al., 2017; Blumenthal et al., 2020). Another review also highlighted that the evidence for the beneficial effects of multidomain interventions is limited (Dedeyne et al., 2017). Future, more rigorously designed studies are needed to compare the effects of multidomain interventions with single-domain interventions, particularly emphasizing comparisons between exercise and cognitive interventions.

Mechanisms through which multidomain interventions exert beneficial effects on cognitive function remain unclear. Current research mainly explains the mechanisms through three aspects: (1) increasing cerebral blood flow, (2) modifying neurotransmitter release, and (3) modifying the structure and functional activities of the central nervous system (Gligoroska & Manchevska, 2012; Serra et al., 2020). Specifically, animal model studies have shown that increased physical exercise can promote hippocampal growth (van Praag et al., 2005), grey matter increase (Sumiyoshi et al., 2014), synaptogenesis (Jiang et al., 2024), and angiogenesis (Pereira et al., 2007), as well as induce the release of brain-derived neurotrophic factor (Almeida et al., 2015) and insulin-like growth factor-1 (Carro et al., 2000). In human studies, physical exercise has also been found to influence grey matter volume (Arenaza-Urquijo et al., 2017), cortical volume (Clark et al., 2019), and hippocampal volume (Ten Brinke et al., 2015), while also promoting the release of brain-derived neurotrophic factor (Marston et al., 2017). Nutritional interventions may help delay cognitive decline and the progression of dementia by improving cell membrane fluidity and vascular endothelial function (Howe et al., 2018; Teixeira et al., 2019), reducing inflammation and oxidative stress (Allès et al., 2012; Miquel et al., 2018; Tang et al., 2015), and promoting neurogenesis and neuronal connectivity (Dauncey, 2009; Gomez-Pinilla & Tyagi, 2013). Regarding cognitive interventions, Kempermann, Kuhn & Gage (1997) demonstrated in a mouse model that enriched environments protect white matter integrity by enhancing neurogenesis (Kempermann, Kuhn & Gage, 1997). Another study observed similar brain volume changes in older adults after cognitive training (Tait et al., 2017).

Another question worth investigating is why the cognitive effects of multidomain interventions are superior to those of single-domain interventions. One possible reason is that nutrition, exercise, and cognitive interventions have complementary and interactive effects. For example, regarding increased cerebral blood flow, consuming foods rich in antioxidants can improve endothelial function and reduce arterial stiffness (Howe et al., 2018), while regular participation in aerobic exercise can slow the decline in cerebral hemodynamics, such as cerebral arterial blood flow velocity and mean arterial pressure, associated with aging (Ainslie et al., 2008; Bailey et al., 2013). These two interventions synergistically improve cerebral blood flow, thereby enhancing nutrient delivery and promoting the release of neurotransmitters that aid cognition. Moreover, there are complementary mechanisms between exercise and cognitive interventions. For example, existing studies have shown that exercise supports the proliferation and division of neuronal cells (Hötting & Röder, 2013; Shors et al., 2012), while cognitive interventions help to promote the survival of these newborn cells (Kempermann, Kuhn & Gage, 1997).

Concerning physical function, the present study indicated that multidomain interventions could not significantly improve physical performance compared to the control group. This contrasts with findings from two review studies that indicated that multidomain interventions could improve physical performance compared to the control group (Giné-Garriga et al., 2015; Liao, Shen & Li, 2023). The possible explanations for our study, in which multidomain interventions did not significantly improve physical function, are that the intervention format in one study was 30 min of home-based exercise (Andrieu et al., 2017), the intervention frequency in the other study was twice a week (Barreto et al., 2021), and the intervention was not supervised. Such designs may not fully capture the benefits of exercise intervention. For other physical outcomes, some review studies demonstrated that multidomain interventions could effectively enhance strength, speed, and stability (Dedeyne et al., 2017; García-Llorente et al., 2024). For example, one review suggested that multidomain interventions were effective in improving muscle strength in (pre)frail elderly compared to single domain interventions (Dedeyne et al., 2017). However, another review found no significant difference in those outcomes when compared to exercise or nutritional interventions alone (Pérez-Bilbao et al., 2023). As a result, more research is needed to investigate the effects of multidomain interventions on physical function.

The American College of Sports Medicine (ACSM) and the American Heart Association (AHA) recommend that older adults engage in moderate-intensity aerobic exercise for 30 min five times per week or vigorous-intensity aerobic exercise for 20 min three times per week. In addition, older adults should perform moderate-intensity resistance training twice weekly. Moreover, ACSM and AHA suggest that incorporating balance and flexibility exercises can provide additional benefits (Nelson et al., 2007). Among the six studies included in this study, four implemented both aerobic and resistance training (Barreto et al., 2018; Chatterjee et al., 2022; Liu et al., 2023; Macpherson et al., 2022), while two focused solely on aerobic exercise (Andrieu et al., 2017; Blumenthal et al., 2020). All included studies met the recommended duration of physical activity, though only one study met the ACSM-recommended frequency of five sessions per week (Andrieu et al., 2017). Given the complexity of multidomain interventions, it is challenging to implement five exercise sessions per week due to the need to include multiple interventions. Notably, the expert consensus guidelines on exercise recommendations for older adults indicate that engaging in physical activity three times a week also shows significant benefits (Izquierdo et al., 2021). Therefore, future multidomain interventions should incorporate exercise sessions at a frequency of three or more times per week.

Strengths and limitations

This study has three main advantages. First, most studies included in the analysis are high-quality RCTs, ensuring the reliability of the intervention effects. Second, strict population inclusion criteria were applied, focusing exclusively on older adults with SCD. This distinguishes this study from most studies that include elderly individuals with varying cognitive function status, potentially introducing bias into the results. Third, this study assessed both cognitive and physical function indicators, providing a comprehensive evaluation of the effects of multidomain interventions.

This study also has three limitations. First, the included studies exhibit considerable variability in the characteristics of the multidomain interventions, such as different types of exercise and nutritional interventions, as well as varying lengths, doses, frequencies, and durations. This heterogeneity may have affected the pooled effect and the generalizability of the results. Secondly, this study limited the scope of the multidomain intervention components, aiming to reduce the heterogeneity by focusing on specific combinations. However, this approach inevitably excludes literature that includes other non-pharmacological interventions (such as psychosocial or sleep interventions), which may lead to a ‘sliced literature’ effect. Lastly, due to the paucity of available data, this study assessed only one physical function and compared the effects of multidomain interventions with single nutritional interventions. The limited number of studies available for this comparison may have compromised the robustness of the findings.

Conclusions

The present study, involving six RCTs, investigated the effects of multidomain interventions on cognitive and physical functions. The results demonstrated that multidomain interventions significantly enhanced executive function and memory compared to the control group and nutritional interventions. This study further validates the effects of multidomain interventions on older adults with SCD and provides valuable insights for both clinical practice and future research. However, the impacts of multidomain interventions on global cognition and physical performance are inconsistent. Future research should focus on more robust multi-armed studies to explore the effects and underlying mechanisms of multidomain interventions.

Supplemental Information