We estimated the risk of cognitive decline based on variation pattern of cholesterol different from previous studies. The results showed that increased TC or NHDL-C from normal to high level was significantly associated with decreased risk of global cognitive decline in females compared with stable normal levels. Decreased TC from high to normal level was significantly associated with increased risk of memory function decline both in males and females compared with stable high levels; but only significant in females for decreased NHDL-C. Persistent high levels of TC and NHDL-C did not significantly increase the risk of any cognitive dimensions. For males, decreased LDL-C from high to normal level was significantly associated with decreased risk of mental intactness declined, and persistent high levels of LDL-C was significantly associated with increased risk of mental intactness declined. The variation of HDL-C was not associated with cognitive decline.
TC and NHDL-C are well-established risk factors for cardiovascular diseases, diabetes and stroke. Multi-center study has determined a significant association of baseline NHDL-C with long-term cardiovascular events, including coronary heart disease, unstable angina, coronary death, coronary revascularization and ischemic stroke [17]. Because the cardiovascular health is helpful in keeping normal cognition, the role of cholesterol on cognition should be in line with cardiovascular disease. However, the association between cholesterol and cognitive function has been divergent in recent years. A study using urban community participants from Shanghai, China indicated that higher LDL-C was associated with low risk of dementia [6]. There were also randomized trials indicated that the therapy of lowering cholesterol by statins diminished cognitive function slightly [18, 19]. Therefore, the effect of cholesterol variability on cognition remains confusing.
Previous studies have been clarified associations of higher visit-to-visit TC, NHDL-C and LDL-C with cognitive decline [10, 12, 20]. In these studies, CV, VIM and SD reflected the degree of dispersion of cholesterol including downward and upward fluctuations. Not only the stability of cholesterol, but also the long-term unidirectional trend was important to cognitive function. But in previous studies, we still don't know which direction of cholesterol variation can protect cognition in long-term observation. Now we found that the variations of TC and NHDL-C from normal to high level had a protective role on cognitive function, which were in inconformity to results from CV, VIM and SD. There was no significant difference in CV, VIM and SD between normal-high and high-normal groups, and the high CV, VIM and SD may also appeared both in normal and abnormal cholesterol range, which led to ambiguity of association between the direction of cholesterol variation and cognitive decline.
Jianian Hua et al. found a negative association between intra-individual SD and memory function in males using the same survey data as our present study [12]. Stable cholesterol is certainly favorable for cognitive ability. But we did found that the association between unidirectional variation of cholesterol and cognitive decline was differently from that CV, VIM or SD. Generally, we only carry out lipid-lowering intervention when blood lipids exceed the ideal levels. So our results added certain guiding significance for whether lipid-lowering intervention.
Several studies support our findings. A study based on Finland population indicated that moderate decreased serum TC from midlife to late life was associated with 3.5 times risk of cognitive impairment [21]. The same conclusion was documented in Sweden females, in which the lowest quartile of TC difference (refers to greatest decrease in TC) after 32 years follow up was associated with 2.37 times risk of dementia [22]. In Japanese male population, the decreased trend of serum TC was occurred at least 15 years before the diagnosis of dementia, and was lower than those without dementia [23]. In a birth cohort study from Germany, participants with incidence of mild cognitive impairment and Alzheimer's disease had decreased TC level from baseline to endpoint but remained stable in those who remained healthy [24]. The above studies suggested that TC was protective for cognitive function which was consistent with our study. After subtracted HDL-C from TC, the same results for NHDL-C still existed. LDL-C is the main component of NHDL-C. However, the variation of LDL-C was negatively associated with cognitive function in our study, although only in males. This suggested that thoughtful consideration of selective regulate lipids for improving cognitive function is needed.
The positive association of TC and NHDL-C with cognitive function may be partly explained by BMI. Higher BMI is one consequence of hypercholesterolemia [25]. In our data, we found a lower baseline BMI among participants with onset of cognitive decline in 2015. The lower BMI may likely to induce a decreased tendency of TC or NHDL-C. Another critical factor affects cognitive function is age. But in our analysis, there was no difference in age among four groups of cholesterol. The cognitive tendency was comparable for each group individuals in terms of age. An interesting finding was that the association between LDL-C and cognitive decline was opposite to that between TC, NHDL-C and cognitive decline, although only for mental intactness in males. This may be explained through various lipoprotein transport mechanisms. LDL transports cholesterol to peripheral tissues, which is more sensitive to the accumulation of peripheral inflammation and lipid plaques; while HDL reverses transports cholesterol from peripheral tissues to liver for recirculation or excretion in the form of bile acid [26]. TC contains LDL-C and HDL-C. The respective association of LDL-C and HDL-C with cognition may be obscured by this synthesis due to the opposite effect of members. NHDL-C contains LDL-C, which demonstrated that non-LDL-C and NHDL-C cholesterols were key betrayal of LDL-C.
The menstruation of females may be a cause of gender-specific for association of cholesterol with cognitive decline. Postmenopausal females have less estrogen, which reduced level of cholesterol ester transfer protein [27]. The presence of menopause makes more violent fluctuation of cholesterol in females than that in males. Indeed, we found the number of females was significant higher than male in normal-high, high-normal and high-high group. Even in premenopausal with a regular menstrual cycle, the effect of estrogen is stronger than that androgen in males. We also adjusted menstrual status in subgroup. Meanwhile, the differences in sample size result from gender-distribution should be noticed. Sample size is a factor that cannot be ignored for population study. Furthermore, in normal-high, high-normal and high-high group, the age of females was older than normal-normal group, which was more sensitive to cognitive decline; but the age of males was younger. This was consistent with that older females were more likely to be postmenopausal, and had a higher level of TC, NHDL-C or LDL-C. Therefore, the female may bear the brunt of dyslipidemia and cognitive impairment.