Improvement of AD risk scores by use of the Aβ1-42/Aβ1-40 ratio

Jan 20, 2020

By Dr. Sandra Langer et al., Fujirebio

 

Previous posts in this series:

Don't miss the upcoming related Insight articles over the next weeks and get deeper information on the value of using Aβ1-42/Aβ1-40 ratio. Or simply download the complete guidance document now! (requires a Premium eServices account)

 

Different scores have been developed to provide an interpretation of biomarker results for AD diagnosis or risk prediction. Here we will give two examples on the impact of Aβ1-42/Aβ1-40 ratio on risk scores.

 
Erlangen Score

  • Risk scores or algorithms (like Erlangen score or PLM score) are improved by implementing the Aβ1-42/Aβ1-40 ratio into the respective algorithm.
  • Use of the Aβ1-42/Aβ1-40 ratio especially reduces discrepancies in biologically doubtful cases and can therefore increase the confidence in the diagnosis.

The Erlangen score uses a risk graph approach. A comparison of the Erlangen score obtained with three (use of only tTau, pTau, and Aβ1-42) and with four biomarkers (use of tTau, pTau, Aβ1-42, and Aβ1-42/40 ratio) as done in the Erlangen score validation study2 clearly shows that the results based on four biomarkers correlate better with the percent of the subjects progressing to AD dementia than the scores calculated with three biomarkers only (compare the light grey versus dark grey bars in fig. 2). Not including the Aβ1-42/40 ratio leads to a counterintuitive result that the proportion of the subjects progressing to AD dementia is larger in the group with a lower score. This might be cautiously considered as another argument favouring the “four-biomarkers” approach, which is also in agreement with other studies confirming better diagnostic performance of the Aβ1-42/1-40 ratio compared to the Aβ1-42 concentration alone.

Figure 2 From: Lewczuk et al. Journal of Alzheimer's Disease 2015: Percentage of the DCN subjects in the MCI stage progressing to AF in the follow-up time (1-4 years). Light grey bars indicate the results when only three biomarkers (Aβ1-42, tTau and pTau181) were considered; dark-grey bars indicate the results when four biomarkers (Aβ1-42, Aβ42/40 ratio, tTau, and pTau181) were considered. In brackets, the total number of patients with a given score is presented in the four-biomarker model.

Figure 2 From: Lewczuk et al. Journal of Alzheimer's Disease 2015: Percentage of the DCN subjects in the MCI stage progressing to AF in the follow-up time (1-4 years). Light grey bars indicate the results when only three biomarkers (Aβ1-42, tTau and pTau181) were considered; dark-grey bars indicate the results when four biomarkers (Aβ1-42, Aβ42/40 ratio, tTau, and pTau181) were considered. In brackets, the total number of patients with a given score is presented in the four-biomarker model.

PLM Scale / PLMR scale

The PLM Scale uses the results of the three AD biomarkers Aβ1-42, tTau and pTau1. Recently, Lehmann et al. published an optimized PLMR score, that integrates the CSF Aβ1-42/Aβ1-40 ratio instead of the Aβ1-42 value alone. In a study with two independent cohorts it could be demonstrated that this PLMR scale could better define AD patients in a clinical routine in a memory centre. The number of discordant biomarker profiles have been significantly reduced by integrating the Aβ1-42/Aβ1-40 ratio into the PLM scale.

Including the Aβ1-42/Aβ1-40 ratio into this tool helps especially in reducing the discrepancies in biologically doubtful cases and can therefore increase the confidence in the diagnosis1.

Non-pathologic biomarkers in PLM scale are defined as below cut-off for Aβ1-42, above cut off for tTau and pTau.

Non-pathologic biomarkers for PLMR scale are defined as below cut-off for Aβ1-42/Aβ1-40 ratio, above cut off for tTau and pTau.

Figure 3 Adapted from: Lehmann et al. Frontiers in Aging Neuroscience 2018 PLM scale and PLMR scale group distributions in one of the two tested cohort (AD and NAD subjects): Scale 0: No pathologic biomarker (predictive value for AD 9.6%), Scale 1: 1 pathologic biomarker out of three (predictive value for AD 24.7%), Scale 2: 2 pathologic biomarkers out of three (predictive value 77.2%), Scale 3: All 3 biomarkers are pathologic (predictive value 94.2%)

Figure 3 Adapted from: Lehmann et al. Frontiers in Aging Neuroscience 2018 PLM scale and PLMR scale group distributions in one of the two tested cohort (AD and NAD subjects): Scale 0: No pathologic biomarker (predictive value for AD 9.6%), Scale 1: 1 pathologic biomarker out of three (predictive value for AD 24.7%), Scale 2: 2 pathologic biomarkers out of three (predictive value 77.2%), Scale 3: All 3 biomarkers are pathologic (predictive value 94.2%)

References

  1. Lehmann, Sylvain, et al. "Relevance of Aβ42/40 ratio for detection of Alzheimer Disease pathology in clinical routine: The PLMR scale." Frontiers in Aging Neuroscience 10 (2018): 138.
  2. Lewczuk, Piotr, et al. "Validation of the Erlangen score algorithm for the prediction of the development of dementia due to Alzheimer’s disease in pre-dementia subjects." Journal of Alzheimer's Disease 48.2 (2015): 433-441.
Close