Triglycerides (TG) are simple lipids involved in the storage and transport of energy. There are two sources for serum TG: gut absorption and liver synthesis. Depending on three fatty acids and their combinations, there are more than 6000 species of TG [23]. These lipids are an important part of triglyceride-rich proteins (TRLs) which comprise very-low-density lipoprotein (VLDL), chylomicrons, and the particles that remain after their catabolism [20,22].

Banks et al. detected TG in human cerebrospinal fluid and moreover, they found that radioactive TG triolein crosses the blood-brain barrier (BBB) in mice. The authors have demonstrated the improvement of cognition after gemfibrozil administration with a subsequent decrease of serum TG level [24]. However, the evidence regarding the passage of the BBB by the TG is very scarce and requires more studies. Data from animal studies suggest that TG contribute to cognitive decline through the impaired maintenance of the N-methyl-d-aspartate component of hippocampal long-term potentiation, and consider that lowering TG could reverse the cognitive impairment in mice [25].

Data from several longitudinal studies have shown a correlation between serum TG in midlife and the risk of cognitive impairment in the elderly [7,26]. Increased TG level in middle age years was associated with the risk of dementia 25 years later in a large cohort including Japanese-American men in the Honolulu–Asia Aging Study [26]. A recent longitudinal cohort study with 13,997 eligible participants pointed out the association of midlife high total cholesterol and TG levels with advanced cognitive decline assessed by memory, executive function, sustained attention, and processing speed test, after 20 years of follow-up [7]. Reynolds et al. found that lower TG levels in women before 65 years (average age at baseline for lipid evaluation was 63.76 years) were associated with and were also predictive for future better memory performance and verbal abilities in a 16-year longitudinal study with 819 participants. The TG concentrations were less predictive in men [27]. Ancelin et al. outlined a gender-dependent outcome of TG levels and incidence of cognitive impairment risk in a 7-year follow-up study with 7053 participants: 4308 women and 2745 men, mean age 73.9 and 73.7 years, respectively. In men without cardiovascular diseases increased TG levels were associated with a significant incidence of all-cause dementia, except for AD. Interestingly, in women, both low and high TG were correlated with decreased risk of AD, independent of apolipoprotein E (APOE) status and vascular factors. The authors suggested that the association of high TG level and low risk of AD could be restricted to those persons carrying the AA polymorphism of apolipoprotein A5 (APOA5) [28].

A few years earlier, Henderson et al. found no association between TG level and memory in a longitudinal study performed on 326 women (aged 52–63 years) during 8 years of follow-up [29]. Moreover, a prospective study with 5 years follow-up indicated an association between increased TG levels and a lower risk of cognitive decline in 930 Chinese subjects with a mean age of 94 years [30].

The association of higher TG levels with poorer cognitive function is also supported by several cross-sectional studies. The outcomes from a case-control Chinese study with 112 MCI subjects and 115 cognitively normal participants aged >65 years indicated a correlation between high TG level and MCI [31]. Data from another cross-sectional study performed on 121 African-American participants (mean age 43.74 years) revealed inferior results of verbal learning performance on the California Verbal Learning Test II (CVLT) in subjects with increased TG levels, without any relation to blood pressure [32]. In addition, higher levels of serum TG were significantly associated with poorer executive function, but not with memory, independent of other vascular risk factors, apolipoprotein E4 (ApoE4) status, and cerebral white matter microstructure. 251 non-demented elderly have participated in this cross-sectional study, mean age 78 years, 54% male [33].

In contrast, the results from a Chinese cross-sectional study performed on 836 subjects (majority females) indicated the association of high normal plasma TG with better cognitive function in elderly over 80 years [34]. Increased TG were also negatively associated with cognitive impairment in male aged 40–55 years in a cross-sectional study with 1762 participants [35]. The results from the cross-sectional Leiden 85-Plus Study showed no correlation of the TG concentrations with cognitive decline in 561 subjects aged ≥85 years [36]. Furthermore, in males, recent data from a cross-sectional study on 2150 Japanese participants (aged 60–90 years) found that high TG levels decrease the global cognitive impairment risk [37].

A recent cross-sectional study with 689 participants from the Alzheimer’s Disease Neuroimaging Initiative cohort (160 with AD, 339 with MCI, and 190 cognitively normal) revealed that reduced levels of two long-chain, polyunsaturated fatty acid-containing TG (PUTG), principal component 3 (PC3) and principal component 5 (PC5), respectively, were associated with lower cognitive performance, hippocampal and entorhinal cortical atrophy [23].

Table 1 summarizes the included studies about triglycerides and cognitive decline.

The etiology of AD remains unclear, but evidence suggests that the major neuropathological hallmark of AD is the accumulation of amyloid protein in senile plaques due to over-production or impaired clearance of β-amyloid (Aβ) peptides and the deposition of neurofibrillary tangles (NFTs), which give rise to synaptic loss and neurodegeneration [38,39]. It has been proposed that asymptomatic cerebral amyloidosis should be the first stage of preclinical AD for research purposes. On the other hand, brain Aβ deposition is noticed in a substantial percentage of elderly without cognitive impairment [40]. Not all individuals with brain Aβ deposition will exhibit dementia symptoms during their life. A possible protective role of genetic factors, brain reserve, or environmental factors is taken into consideration [41]. The biomarkers of brain Aβ plaque load are positron emission tomography (PET) Aβ imaging and low cerebrospinal fluid (CSF) Aβ42 [14]. Nevertheless, these markers appear to assess the fibrillar forms of Aβ but not the oligomeric ones which seem to be critical for synaptic impairment [41].

Regarding the dynamic of brain amyloid-beta deposition, there is a sigmoid shape of the evolution with an exponential phase that corresponds to normal cognitive status, and a plateau phase which is reached before the appearance of clinical symptoms or to atrophy on MRI [14]. Considering that cerebral amyloid accumulation is starting at least 10 years before late MCI and AD dementia development, a cross-sectional study might have difficulties identifying contributing factors to cerebral amyloidosis in participants with cognitive deterioration [42]. A recent study performed on 942 elderly individuals (average age 79.7 years) in the Mayo Clinic Study of Aging revealed an association between midlife dyslipidemia (total cholesterol) and amyloid β brain deposition [43]. Nägga et al. support the idea that risk factors related to early AD changes should be investigated in normal cognitive participants [44].

Data from preclinical studies revealed that in AD mouse models with abundant plasma Aβ, VLDL (very low-density lipoprotein) TG levels precede amyloid brain deposition [45].

In humans, a cross-sectional study showed that high serum TG level in normal cognitive individuals (mean age 70.2 ± 5.7 years) was associated with more global Aβ PET deposition (cerebral amyloidosis) after APOE4 adjustment, while no correlation was found for total cholesterol, HDL-C or LDL-C. However, in this study amyloid-beta deposition in medial temporal, occipital, and basal ganglia regions was not positively associated with high serum TG [42].

The results of another longitudinal cohort study of 318 cognitively normal individuals concluded that increased midlife triglycerides (mean age 54 years) were associated with abnormal CSF Aβ42 together with CSF Aβ42/p-tau ratio 20 years later, after adjusting for multiple vascular factors, education, age, and APOE4 [44].

Data from Framingham Heart Study outline the influence of hypertriglyceridemia (≥287 mg/dL for males and ≥226 mg/dL for females) during midlife (40–60 years of age) on late-onset AD risk in APOE e4 negative participants, after adjustment for systolic blood pressure, and based on genetic markers. This longitudinal study with over 10 years of follow-up included 157 cases and 2882 controls with AD status and genotypes available [46]. This information can be used to tailor preventive strategies targeting triglyceride levels.

However, evidence is still controversial regarding the association of increased serum TG levels and cerebral amyloidosis or AD. A small cross-sectional study performed on 74 individuals (3 with mild dementia, 38 with mild cognitive impairment, and 33 clinically normal) did not find a correlation between TG or total cholesterol levels and cerebral Aβ quantified by Global PIB Index. Still, there was an independent association between LDL cholesterol, HDL cholesterol, and amyloid brain deposition [47].

Proitsi et al. found no difference between AD patients and controls regarding serum TG, total cholesterol, LDL, and HDL cholesterol in a subgroup of participants (102 cases versus 104 controls). However, they underlined the association of low-chain and very-low-chain triglycerides with AD in the untargeted lipidomic analysis on 142 AD and 152 control subjects [48]. The results of a large study using Mendelian randomization were negative regarding any causal association between lipid fractions and AD risk. The study included 17,008 participants with AD and 37,154 controls [49]. Evidence from cross-sectional small sample studies has shown a decreased TG level in dementia subjects [50,51]. The statistically significant values were only in Lepara et al. study that has outlined the presence of lower TG levels in probable AD individuals (24 females and 6 males) compared to control groups [50].

The relationship between TG and cerebral amyloidosis may be related to the circulating complex of amyloid β synthesized in enterocytes and triglyceride-rich lipoproteins that could disrupt the blood-brain barrier and subsequently increase brain amyloid deposition [52]. Furthermore, another possible pathway that leads to increased TG and amyloid-beta accumulation is the absence of peroxisome proliferator-activated receptor-gamma (PPAR-r) [53]. Studies on animals suggest that this receptor regulates adipose triglyceride lipase and facilitates the clearance of brain β amyloid [54].

In Table 2, we have summarized the main studies about triglycerides and cerebral amyloidosis.

Vascular cognitive impairment (VCI) describes a large spectrum of cognitive dysfunction caused by vascular diseases including stroke, silent brain infarction, and subclinical brain injury [4]. The incidence and prevalence of VCI are age-related [55]. In individuals with dementia, there is frequently evidence of the coexistence of both vascular and AD lesions making it difficult to assess the exact contribution of each one to the cognitive decline [5,9,55].

In a study performed on 1143 subjects, the degree of cerebral atherosclerosis and arteriosclerosis was correlated with the severity of the cognitive impairment, including perceptual speed, which indicates a vascular cause, and episodic memory, which represents an important feature of AD [6]. The main incriminated mechanism of cognitive decline by cerebrovascular disease is represented by hypoperfusion [5]. Besides the well-known involvement of hypercholesterolemia in the atherosclerosis process, there is evidence that through endothelial dysfunction, foam cell formation, inflammation, and cytokines regulation, TG are participating in the atherosclerotic cardiovascular disease risk [19,20,21,22].

Studies have shown that the constellation of metabolic abnormalities represented by metabolic syndrome (MetS) is associated with a higher risk of vascular dementia but not AD [26,56,57]. In elderly Americans (≥60 years), MetS has a prevalence of 54.9 ± 1.7% [58].

Metabolic syndrome represents a cluster of conditions: abdominal obesity (waist circumference >40 inches in men and >35 inches in women), hyperglycemia (>100 mg/dL)/pharmacological treatment, high TG (≥150 mg/dL or pharmacological treatment), low HDL cholesterol (<40 mg/dL in men and <50 mg/dL—female or pharmacological treatment) and hypertension (>130/85 mmHg or pharmacological treatment). According to the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) revised criteria, MetS diagnostic requires at least three of these five cardiometabolic parameters [59]. It was found that MetS may increase vascular dementia risk, but the exact mechanism is not known. A possible explanation could be related to the microvascular damage which leads to white matter deterioration and disturbance of neuronal connectivity [57]. Several studies outlined the risk of cognitive impairment in people with white matter lesions [60,61].

MetS increases the risk of leukoaraiosis [62], silent brain infarction [63], and clinical stroke [64]. Impaired fasting glucose and hypertriglyceridemia were associated with leukoaraiosis independently of elevated blood pressure in a Japanese population with 1030 healthy subjects (mean age 52.7 years; 534 men and 496 women) [62]. However, impaired glucose tolerance alone was found to be an equal predictor of stroke events in elderly individuals as MetS [64]. MetS was found to be significantly associated with silent brain infarction, a predictor of both clinical overt stroke and dementia [63].

Regarding the risk of vascular dementia and MetS components, Solfrizzi et al. observed a synergistic effect with statistical significance of all items compared to the additive risk of individual components. The link was even stronger in subjects with high inflammation status and after the exclusion of undernourished participants. The study was performed on 2097 participants with 3.5 years of follow-up [57]. On the other hand, Raffaitin et al. demonstrated an independent association of each component of MetS with the risk of dementia. The cohort was a subsample of the Three-City Study and included 7087 participants aged ≥65 years with 4 years of follow-up [56]. High TG level was the only component of metabolic syndrome that was significantly associated with the incidence of all-cause and vascular dementia, while diabetes, but not impaired fasting glycemia, was significantly associated with all-cause and vascular dementia [56,65].

Bowler emphasizes the importance of recognizing cardiovascular risk factors, comprising all components of MetS, as risk factors for vascular dementia [66]. Considering single components of MetS and the risk of silent brain infarction, a study with 1588 healthy subjects determined that only high blood pressure and impaired fasting glucose had a strong significance, while high TG and low HDL levels showed marginal significance [63].

A cross-sectional study performed on 202 non-demented participants of the Biomarker Development for Postoperative Cognitive Impairment in the Elderly (BioCog study, patients aged 65-87 years with elective surgery in centres from Utrecht, the Netherlands, and Berlin, Germany) supports the idea that high TG level increases twice the possibility of cognitive impairment. However, this association was no longer statistically significant after adjustment for cerebrovascular and coronary heart disease [67]. A recent cross-sectional study with 108 participants aged ≥60 years with memory complaints from the Australian Imaging Biomarkers and Lifestyle (AIBL) study showed a positive correlation of the multiple risk factors in MetS with lower executive function and global cognitive performance [68].

No association was found between MetS and AD [56,57]. In contrast, a cross-sectional study performed on a random population-based sample of 980 individuals has shown that the prevalence of AD was higher in women with MetS, but not in men. Furthermore, over 80% of women with AD had MetS [69].

Table 3 summarizes the main studies about triglycerides and the vascular cognitive impairment.