Plasma p-tau181 as an outcome and predictor of multidomain intervention effects: a secondary analysis of a randomised, controlled, dementia prevention trial

Background It is unknown whether multidomain interventions, which might preserve late-life cognition, affect Alzheimer's disease pathology. Previous studies measured cerebrospinal fluid and imaging Alzheimer's disease biomarkers in small subsamples of multidomain trial participants. Newly developed assays enable the measurement of blood-based Alzheimer's disease biomarkers in larger samples. We aimed to assess whether plasma tau phosphorylated at threonine 181 (p-tau181) was able to detect or predict 3-year multidomain intervention effects. Methods This is a secondary analysis of the randomised, controlled, Multidomain Alzheimer Prevention Trial (MAPT) testing a 3-year multidomain intervention, omega-3 fatty acid supplementation, or both versus placebo, in individuals aged 70 years and older in 13 memory centres in France and Monaco. Plasma p-tau181 was measured in stored blood samples in a subsample of 527 participants on an intention-to-treat basis. Changes in cognitive score were calculated as a composite measure using the average of Z scores for the following tests: Mini Mental State Examination orientation items, Free and Cued Selective Reminding Test (sum of free and total recall scores), category fluency, and Digit Symbol Substitution Test. Intervention effects on 3-year change in p-tau181 concentration were estimated by use of a linear mixed model with centre-specific random intercepts. Findings Recruitment took place between May 30, 2008, and Feb 24, 2011. Median baseline plasma p-tau181 was 8<middle dot>8 pg/mL (IQR 6<middle dot>7-11<middle dot>9) in the total sample, and significantly higher in older individuals, men, APOE epsilon 4 carriers, and participants with renal dysfunction or a positive PET amyloid scan. During 3-year follow-up, individuals with raised baseline p-tau181 underwent greater cognitive decline (eg, mean difference in 3-year change on the composite cognitive score between control group participants with normal and abnormal baseline levels of p-tau was -0<middle dot>34 [effect size -0<middle dot>52; 95% CI -0<middle dot>61 to 0<middle dot>07] in the fully adjusted model using a 12<middle dot>4 pg/mL cutoff for abnormal baseline p-tau181), but there were no intervention effects on change in p-tau181 either in this subgroup or the total population, and no effect on cognitive change in individuals with raised baseline p-tau181 (eg, in the fully adjusted model using the 12<middle dot>4 pg/mL cutoff for p-tau181 abnormality, the mean difference [95% CI] in this subgroup in 3-year decline on the composite cognitive score between the control group and the multidomain + omega-3 group, the omega-3 group, and the multidomain intervention group, was, respectively: 0<middle dot>13 [-0<middle dot>21 to 0<middle dot>47], 0<middle dot>03 [-0<middle dot>30 to 0<middle dot>36], and 0<middle dot>10 [-0<middle dot>26 to 0<middle dot>46]). Surprisingly, individuals with raised baseline p-tau181 showed a decrease in p-tau181 during follow-up (eg, unadjusted mean [95% CI] 3-year change was -3<middle dot>01 pg/mL (-4<middle dot>45 to -1<middle dot>56) in control group subjects with abnormal baseline p-tau181 [using the 12<middle dot>4 pg/mL abnormal p-tau cutoff]). Interpretation Our results support the utility of p-tau181 as a prognostic biomarker, but it did not predict or detect intervention effects in this study. Further investigation of its usefulness as a prevention trial outcome measure is required.