Abstract
Background: Both the genetic and pathological studies link Alzheimer's disease (AD) to the triggering receptor expressed on myeloid cells 2 (TREM2). A large number of studies have explored the value of cerebrospinal fluid (CSF) soluble TREM2 (sTREM2) levels as a biomarker for the diagnosis and prediction of AD; however, the findings are inconsistent. We aimed to review the studies that investigated the association of CSF sTREM2 levels and AD risk, and to provide the recommendations for future research. Methods and Results: A systematic literature search was performed using the MEDLINE, EMBASE, and Web of Science (all databases) databases. The meta-analysis for the association between the CSF sTREM2 levels and AD risk included 15 studies (17 comparisons) with a total of 1,153 cases and 1,626 controls. The total results showed that the higher CSF sTREM2 levels and AD risk were associated [standardized mean difference (SMD) = 0.428, 95% CI (0.213, 0.643), I (2) = 81.1%]. However, the analysis of the subgroup of "age difference ≤ 2 years" indicated that sTREM2 was not associated with AD [SMD = 0.090, 95% CI (-0.092, 0.272), I(2) = 27.4%] and had a significantly lower heterogeneity. Combining the results of the "age difference of 5-10 years" [SMD = 0.497, 95% CI (0.139, 0.855), I (2) = 82.5%] and "age difference > 10 years" [SMD = 0.747, 95% CI (0.472, 1.023), I (2) = 50.0%] subgroups showed that the difference in CSF sTREM2 between the AD and control groups was positively correlated with the age difference. A meta-regression analysis showed that the age difference can explain 33.4% of the between-study variance. By conducting further subgroup analyses of the five age-matched studies (495 cases and 364 controls) according to the measurement method, and whether inclusion criteria containing the requirement for pathological evidence of AD, no changes were observed in the corresponding pooled SMD or in the significance of the results. The meta-analysis result of "age difference ≤ 2 years" group was robust in the sensitivity analysis. Conclusion: The available high-quality evidence does not yet support an association between the CSF sTREM2 levels and AD risk. Age matching between the patients with AD and cognitively unimpaired controls was a major influencing factor in the results.