This meta-analysis explored and summarized the inconsistent data of previous studies investigating the correlation of inflammatory biomarkers with cognitive function and glycemic and lipid profiles in patients with T2DM. From the data of 32 studies including 7483 T2DM patients, the study illustrated that serum IL-6, CRP, and TNF-α levels and MoCA scores showed a moderate negative correlation, IL-6 and IL-8 levels and MMSE scores had a low negative correlation. Previous studies and meta-analyses suggested that T2DM patients with cognitive impairment have significantly increased levels of IL-6, CRP, TNF-α, and IL-8 compared to controls [3, 7, 47]. The results of the present study further supported that inflammatory cytokines may play a fundamental role in the accelerated development of cognitive impairment in T2DM patients. Our findings have also supported that there is a moderate and positive correlation between serum CRP and TNF-α levels and HbA1c, whereas IL-1β and IL-18 levels and MMSE scores showed a lower positive correlation. Further, serum TNF-α level was positively correlated with FBG, serum CRP level was positively correlated with TC, and serum IL-6 level was positively correlated with TG, and effect sizes were all low in strength. However, such findings should be interpreted with caution, given the high heterogeneity of studies.
IL-6 is a potential marker associated with cognitive impairment and diabetes [3]. The results showed that IL-6 was significantly negatively correlated with MMSE (r = -0.352, P < 0.001) and MoCA (r = -0.540, P = 0.006), indicating that the correlation between IL-6 and MoCA was better than that of MMSE. The MoCA is a simple, stand-alone, and sensitive cognitive screening tool for MCI [53]. A meta-analysis of 175 studies in AD patients suggested that peripheral IL-6 may clinically correlate with the severity of AD [4]. These results suggest that IL-6 may become an inflammatory biomarker of T2DM with MCI, but the diagnostic level of T2DM with MCI has not been reported, and further clinical research is needed.
The results showed a significant positive correlation between IL-6 and cognitive function in T2DM patients. Our finding was in accordance with previous study, it revealed that higher plasma IL-6 level was associated with worse performance of executive function and abstract reasoning ability [54]. In both preclinical and clinical investigations, IL-6 has received substantial attention as a pro-inflammatory cytokine with the potential to hasten ongoing neurodegenerative processes in AD [55]. Inflammation could have both beneficial and detrimental effects depending on the severity of the cognitive impairment. During MCI pathology, microglia and astrocytes, key cellular drivers and regulators of neuroinflammation, are involved in Aβ clearance via chemotactic mechanisms, so the activation is beneficial. However, as the disease progresses, activated microglia cause neurodegeneration in the surrounding brain regions by overexpressing pro-inflammatory cytokines like IL-6, which could potentially accelerate the pathological cascade of AD. IL-6 is overexpressed near the Aβ plaques and neurofibrillary tangles (NFT) and is known to increase Aβ peptides production [55, 56]. It has been demonstrated that IL-6 aggravates the neuroinflammatory process by causing the deposition of Aβ plaques, resulting in cytotoxic effects [57]. IL-6 also influences the proliferation and differentiation of central nervous system (CNS) cells via mediating immune responses and inflammatory reactions [18]. The evidence was therefore supported the current meta-analysis.
CRP is a low-grade systemic inflammatory marker as well as a well-known acute phase protein produced by the liver [3, 58]. The expression of CRP is in response to inflammation and is mainly regulated by IL-6 in the acute phase, which plays a causative role in the pathogenesis of T2DM and is associated with the risk of DM complications [18]. Increased peripheral concentration of CRP was also found in patients with AD compared with the control [5]. In agreement with these results, our study reported a moderate negative correlation between CRP levels and MoCA scores in T2DM patients. In addition, CRP has been reported to be associated with atherosclerosis and cardiovascular risk, which could affect white matter (WM) structure through endothelial dysfunction [59]. Epidemiologic and experimental researches have revealed that both atherosclerosis and vascular disease contribute to cognitive impairments [58, 60]. In agreement with the latest study, higher baseline levels of plasma IL-6 and CRP were associated with subsequent cognitive decline in older people with T2DM. Specifically, elevated CRP level was associated with a greater decline in processing speed [54].
TNF-α, a pro-inflammatory cytokine, plays an important role in the pathophysiology of many inflammatory disease states, such as T2DM, AD, and obesity, by regulating inflammatory and immune responses [4, 41]. Our meta-analysis showed a moderate negative correlation between TNF-α level and MMSE/MoCA scores. This finding is consistent with previous investigations have discovered elevated levels of serum TNF-α to be associated with poorer performance on the Paired Associates Learning (PAL) task, which requires significant working memory and involves brain regions such as the prefrontal cortex, medial temporal lobe, hippocampus, basal ganglia and parietal cortex [61]. Another large population study conducted in Scotland analyzed circulating levels of the inflammatory markers including TNF-α, CRP, and IL-6 and their association with cognitive ability in 1066 patients aged 60–75 years with T2DM. It has proven that increased levels of plasma TNF-α, CRP, and IL-6 were significantly associated with poorer general cognitive abilities after adjustment for age and sex [62]. These findings suggest that elevated TNF-α may be particularly critical to T2DM-related cognitive impairment and holds great promise as a diagnostic marker for cognitive impairment in T2DM patients along with other markers.
A low degree of negative correlation between IL-8 and MMSE scores was also observed. In AD patients, peripheral IL-8 level was higher compared with the healthy elderly [63], other meta-analysis have observed lower level of IL-8 in the periphery in MCI compared with the control [5]. These findings suggest that IL-8 is associated with the pathogenesis of cognitive decline and could be used as a potential biomarker to monitor disease progression. IL-8 not only triggers and initiates acute inflammation, recent studies have demonstrated that elevated IL-8 is also involved in the chronic neuro-inflammation in AD. It has been reported that IL-8 may be related to cerebrovascular damage in cognitively impaired patients [64].
HbA1c, haemoglobin with an attached sugar moiety, provides an estimate of an individual’s blood sugar levels over the past three months and is commonly used to assess long-term blood glucose concentration. According to the American Diabetes Association (ADA) diagnostic criteria (2018), a level of HbA1c ≥ 6.5% was used to help define newly diagnosed diabetes [65]. Recent investigations have found the interaction of inflammation with glycemic control in patients with significantly higher levels of HbA1c and a presumably increased risk for cognitive impairment. Insulin and oral diabetic drugs are both known to have anti-inflammatory effects, and their combination may increase this property [9]. These finding could be supported by the fact that there is a positive correlation between serum CRP, TNF-α, IL-1β, and IL-18 levels and HbA1c in the present study. Furthermore, we discovered that serum TNF-α level was positively correlated with FBG, but the correlation was low. Excess glucose causes the major HbA1c to release iron, resulting in free radicals. The BBB permeability is increased by iron overload and excessive free radical production [43]. Geng et al. [66] used the 8-week streptozotocin (STZ) diabetic rat demonstrated that diabetes might increase TNF-α levels, triggering disruption of BBB in the brain, and ultimately leading to cognitive impairment. Similarly, Gorska-Ciebiada et al. [18] have demonstrated a positive relationship between HbAlc level and TNF-α in patients with T2DM. HbAlc, on the other hand, does not require fasting and is less likely to be affected by occasional daily activities than FBG as an indicator of blood glucose variability [67]. Therefore, HbA1c is conveniently stabilized and increasingly used in clinical practice.
The lipid profile is a significant risk factor for cognitive impairment. Individuals with MCI had higher plasma lipid levels than those with normal cognition [68]. A cross-sectional study conducted in China showed that high serum TC and LDL-C levels may be risk factors for cognitive impairment in elderly male and female subjects, respectively. A previous study also found that high serum TC level was associated with an increased risk of AD in elderly participants but only in middle-aged or male participants [69]. Conflicting results have been found in studies relating high TG level with AD. Several studies demonstrated that high TG was inversely correlated with cognitive impairment in middle-aged male subjects, but there were also studies that showed no significant association between TG levels and cognitive impairment in older subjects [68, 70]. Furthermore, dyslipidemia is defined by elevated TC, TG, and LDL-C and low HDL-C. Dyslipidemia is one of the modifiable risk factors for vascular disease, and vascular disease may have an important impact on cognitive impairment [70]. The present meta-analysis confirmed that serum IL-6 level was positively correlated with TG and serum CRP level was positively correlated with TC, suggesting that the inflammatory cytokines IL-6 and CRP may play an intermediary role in this process. Additional large longitudinal studies and large population studies with comprehensive measurements of inflammatory cytokines, lipid profiles, and cognitive function are required.
Several issues may limit the current analysis. First, though most of the results show high heterogeneity, subgroup and sensitivity analyses failed to identify the sources of that heterogeneity, implying that other sources of heterogeneity could not be assessed systematically among the included studies. It also suggests that other factors may also play roles, or that the studies used noticeably different methodologies. For example, studies included in this meta-analysis lack studies excluding or reporting anti-inflammatory medication use, acute infections, comorbidities, or other inflammatory conditions, which could be a source of potential bias identified among the included studies. Second, the present meta-analysis was also limited by the fewer number of effect sizes included in some qualitative synthesis, such as the correlation of TNF-α and TC. By meta-analysis, it did not reach significance. These inflammatory biomarkers and their correlation with glycemic and lipid profiles might be investigated further. Third, because the included studies were observational and mostly cross-sectional, only correlations could be evaluated. More longitudinal studies producing multiperiod data of inflammatory markers, glycemic and lipid profiles, as well as the widely recognized neurodegenerative biomarkers are important for understanding the pathophysiology of cognitive decline in T2DM patients. Furthermore, further studies are needed to determine how these peripheral markers connect with the known markers such as Aβ and tau.