Sample characteristics
Table 2 shows the demographic, pathophysiological and clinical characteristics of the cohort by CSF profile. There was no significant difference between the groups in gender, age, education or CSF measures. The CSF AD profile group showed significantly worse performance (lower scores) on the MMSE test (global cognition), RAVLT and Story tests measuring verbal episodic memory, ROCF immediate recall (Non-verbal episodic memory) and Verbal fluency animal (language) tests compared to the non-AD group (p < 0.05).
Table 2
Participant demographics, CSF marker levels and neuropsychological test scores by CSF profiles
| | | CSF profile | |
| | | Non-AD T-tau/ Aβ42 ≤ 0.52 (n = 24) | AD T-tau/ Aβ42 > 0.52 (n = 28) | p value ᵃ |
Demographics | | | |
| | Gender (M/F) | 16/8 | 17/11 | 0.66 |
| | Age, years | 67 (46–80) | 70 (51–84) | 0.17 |
| | Education, years | 14.0 (9–20) | 12.5 (6–17) | 0.11 |
CSF measures | | | |
| | Aβ42 (pg/ml) | 703 (374–2332) | 454 (160–822) | N/A |
| | T-tau (pg/ml) | 173 (100–722) | 416 (132–838) | N/A |
| | NFL (ng/ml) | 1.9 (0.9–6.5) | 2.5 (1.2–4.5) | 0.15 |
| | YKL-40 (ng/ml) | 165 (83–399) | 203 (124–367) | 0.12 |
| | S100B (pg/ml) | 215 (132–335) | 230 (130–458) | 0.17 |
| | GFAP (ng/ml) | 1.0 (0.1–7.1) | 1.3 (0.5–21.3) | 0.09 |
Cognitive domains | | | |
| Global cognition | | | |
| | MMSE, score | 28 (24–30) | 27 (24–30) | 0.01 |
| Verbal episodic memory | | | |
| | RAVLT immediate recall, score | 36 (23–66) | 26.5 (13–51) | 0.003 |
| | RAVLT delayed recall, score | 6.5 (0–15) | 1.5 (0–12) | < 0.001 |
| | RAVLT recognition-fp, score | 9.0 (3–15) | 5.5 (-3-15) | 0.003 |
| | Story immediate recall, score | 13.5 (5–17) | 8 (1–18) | 0.005 |
| | Story delayed recall, score | 12.0 (1–19) | 5.5 (0–16) | 0.002 |
| Non-verbal episodic memory | | | |
| | ROCF immediate recall, score | 13.3 (0–27) | 7.3 (0–26) | 0.04 |
| | ROCF delayed recall, score | 12.8 (0–25) | 8.5 (0–26) | 0.07 |
| Language | | | |
| | Verbal fluency animal, score | 20 (8–33) | 14 (4–27) | 0.02 |
| | Verbal fluency H + S, score | 24.0 (14–48) | 25.5 (6–63) | 1.00 |
| Processing speed | | | |
| | TMT-A, sec. | 43.5 (21–133) | 48.0 (27–116) | 0.22 |
| | Stroop – part I, sec. | 23.5 (20–42) | 24.5 (17–34) | 0.64 |
| Executive functions | | | |
| | TMT-B, sec. | 109 (44–340) | 153 (60–343) | 0.06 |
| | DSST, score | 8.5 (3–51) | 7.0 (2–16) | 0.24 |
| | Stroop 4th /3rd part, sec. | 2.1 (1.4-4.0) | 2.1 (1.6–5.8) | 0.26 |
Abbreviations: AD Alzheimer’s disease, CSF Cerebrospinal fluid, DDST Digit symbol substitution test, fp false positives, MMSE |
Mini-Mental State – Examination, RAVLT Rey Auditory-Verbal Learning Test, ROCF Rey–Osterrieth complex figure, TMT Trail |
Making Test, N/A Not applicable |
Values are shown as median (range) or as numbers per group, ᵃMann-Whitney U non-parametric test used for continuous variables and Chi-Square test for the categorical variable (gender), p-values not applicable for Aβ42 and T-tau due to their values used for defining CSF profiles |
Diagnostic accuracy of CSF markers and cognitive domains distinguishing between CSF AD vs. non-AD profiles
Accuracies for distinguishing between CSF AD vs. non-AD profiles were based on ROC curves (Table 3). Neuropsychological tests reflecting verbal episodic memory had the highest accuracy compared to other measurements, with all AUCs over 0.70 which is considered fair [47]. The composite z-score (AUC = 0.80, CI: 0.69–0.92) and RAVLT delayed recall (AUC = 0.80, CI: 0.68–0.93) both distinguished the best between the CSF profile groups. Composite z-scores and tests recflecting other cognitive domains all had AUCs below 0.70. AUC for CSF measures ranged from 0.61–0.64, with lower limit of each confidence interval below the value of 0.5.
Table 3
ROC curves – distinguishing between AD and non-AD CSF profiles
| | | AUC | 95% CI (AUC)* |
CSF measuresᵃ | | |
| | GFAP (ng/ml) | 0.64 | 0.48–0.79 |
| | YKL-40 (ng/ml) | 0.63 | 0.47–0.78 |
| | NFL (ng/ml) | 0.62 | 0.45–0.78 |
| | S100B (pg/ml) | 0.61 | 0.46–0.77 |
Cognitive domains | | |
| Verbal episodic memory | | |
| | Composite z-score | 0.80 | 0.69–0.92 |
| | RAVLT delayed recall, score | 0.80 | 0.68–0.93 |
| | Story delayed recall, score | 0.75 | 0.62–0.89 |
| | RAVLT immediate recall, score | 0.74 | 0.61–0.88 |
| | RAVLT recognition-fp, score | 0.74 | 0.61–0.87 |
| | Story immediate recall, score | 0.73 | 0.59–0.86 |
| Non-verbal episodic memory | | |
| | Composite z-score | 0.65 | 0.50–0.81 |
| | ROCF immediate recall, score | 0.66 | 0.51–0.81 |
| | ROCF delayed recall, score | 0.65 | 0.49–0.80 |
| Executive functions | | |
| | Composite z-score | 0.64 | 0.49–0.80 |
| | TMT-B, sec.ᵃ | 0.66 | 0.50–0.81 |
| | DSST, scoreᵃ | 0.60 | 0.44–0.75 |
| | Stroop 4th /3rd part, sec. ᵃ | 0.59 | 0.43–0.75 |
| Language | | |
| | Composite z-score | 0.60 | 0.44–0.76 |
| | Verbal fluency animals, score | 0.68 | 0.54–0.83 |
| | Verbal fluency H + S, score | 0.50 | 0.34–0.66 |
| Processing speed | | |
| | Composite z-score | 0.56 | 0.39–0.72 |
| | TMT-A, sec. ᵃ | 0.60 | 0.44–0.76 |
| | Stroop test – part I, sec. ᵃ | 0.54 | 0.38–0.70 |
Abbreviations: AD Alzheimer’s disease, AUC Area under curve, CI Confidence Intervals, CSF Cerebrospinal fluid, |
DDST Digit symbol substitution test, fp false positives, RAVLT Rey Auditory-Verbal Learning Test, |
ROCF Rey–Osterrieth complex figure, SE Sensitivity, SP Specificity, TMT Trail Making Test |
AUC is the probability that a randomly selected pair of subjects from each CSF profile group is correctly classified, |
*Confidence intervals calculated with DeLong method, ᵅValues are natural log-transformed |
Fig. 2 illustrates the ROC curves for the two cognitive domains and the CSF measure with the highest AUC from Table 3. Verbal episodic memory (AUC=0.80) was superior in distinguishing between CSF AD vs. non-AD profiles compared to non-verbal episodic memory (AUC=0.65) and CSF GFAP (0.64).
Correlations between CSF markers and cognitive domains
Pearson‘s correlations between the CSF markers and the cognitive domains within the whole cohort and among subjects with CSF AD profile are presented in Fig. 3a and 3b, respectively. Z-scores for cognitive domains were adjusted for age and education to control for confounding effects. Within the whole cohort, levels of inflammatory markers YKL-40 (NFL: r = 0.62, p < 0.001; T-tau: 0.46, p = 0.001) and S100B (NFL: r = 0.52, p < 0.001: T-tau: r = 0.43, p = 0.002) correlated significantly with neurodegeneration markers NFL and T-tau. None of the inflammatory markers correlated with Aβ42 levels (p > 0.05). Higher CSF T-tau levels correlated with worse performance on verbal episodic memory (r=-0.28, p < 0.04) and higher CSF GFAP levels with worse performance on executive functions (r=-0.37, p = 0.007). Among those with CSF AD profile, higher levels of CSF NFL correlated with worse performance on verbal episodic memory (r=-0.43, p = 0.02), higher levels of CSF S100B with worse performance on processing speed (r=-0.45, p = 0.02) and higher levels of GFAP levels with worse performances on processing speed (r=-0.68, p < 0.001) and executive functions (r=-0.39, p = 0.04).
Ridge regression estimates for association between CSF marker levels and cognitive domains
Ridge regression was also performed for the estimation of the relationships between CSF measures and each cognitive domain within the whole cohort when all CSF measures were used as predictors (Table 4). Ridge regression is a penalized approach to linear multiple regression, especially useful when dealing with multicollinearity (highly correlated predictors). The method shrinks the slope coefficients towards zero as a consequence of penalization but keeps all the predictors in the model. The order of magnitude of the CSF standardized slope coefficients (st. β) and pearsons‘r coefficients (Fig. 3a) were similar for non-verbal episodic memory, language, processing speed and executive functions. The CSF measure with the highest coefficient from the ridge regression was also the highest one from the Pearson‘s analysis for each of those cognitive domains. The order, on the other hand, differed for verbal episodic memory. Results from ridge regression placed CSF YKL-40 as having the highest coefficient, positively associating with verbal episodic memory (st. β = 0.48), while CSF NFL (st. β =-0.38) and T-tau (st. β =-0.24) ranked second and third, albeit as negative association with that same domain. The combination of NFL and T-tau with YKL-40 (by calculating ratios) was therefore also tested further.
Table 4 Ridge regression estimates for association between CSF marker levels and
cognitive domains within the whole cohort (n = 52)
| | | Cognitive domains - composite z-scoresb |
| | | Verbal episodic memory | Non-verbal episodic memory | Language | Processing speed | Executive functions |
| CSF measuresᵅ | | | | | |
| | Aβ42 (pg/ml) | 0.13 | 0.19 | -0.04 | 0.06 | 0.05 |
| | T-tau (pg/ml) | -0.24 | <0.01 | < 0.01 | 0.09 | 0.01 |
| | NFL (ng/ml) | -0.38 | -0.09 | -0.12 | -0.07 | 0.01 |
| | YKL-40 (ng/ml) | 0.48 | 0.09 | 0.03 | <0.01 | 0.04 |
| | S100B (pg/ml) | 0.04 | -0.14 | -0.11 | -0.09 | -0.06 |
| | GFAP (ng/ml) | -0.06 | -0.01 | -0.01 | -0.12 | -0.17 |
Numbers represent standardized beta coefficients (st. β) |
ᵅValues are natural log-transformed, bAnalyses are adjusted for age and education |
Correlations between CSF markers and cognitive domains by CSF AD profile
Relationships between the CSF measures and cognitive domains with significant Pearson‘s coefficients (p < 0.05), either within the whole cohort or among those with CSF AD profile, are presented in Fig. 4. CSF NFL and T-tau levels were explored as a ratio of YKL-40 levels in relation to verbal episodic memory based on the results from the ridge regression analysis (Table 4). The CSF NFL/YKL-40 ratio showed stronger correlation with verbal episodic memory (r=-0.51, p < 0.001, Fig. 4a) compared to the correlations of the proteins alone ( NFL: r=-0.26, p = 0.06; YKL-40: r = 0.18, p = 0.20). Corresponding analysis based on the CSF AD profiles (Fig. 4b) also revealed stronger correlation between CSF NFL/YKL-40 ratio and the cognitive domain among those with CSF AD profile (r=-0.67, p < 0.001) than without (r=-0.46, p = 0.03). The relationship of T-tau/YKL-40 ratio with verbal episodic memory (r=-0.44, p = 0.001, Fig. 4c) was significant but weaker compared to the NFL/YKL-40 ratio within the whole cohort as well as among the CSF AD profile group (r=-0.35, p = 0.07, Fig. 4d). Correlations between individual protein levels (NFL, YKL-40, T-tau) and verbal episodic memory, both within the whole cohort and by CSF profile, are presented in Additional file 1, S1a-f. Correlations between the NFL/YKL-40 ratio and individual neuropsychological tests reflecting verbal episodic memory are presented in Additional file 1, S2a-e.
Weak negative correlation was found between CSF GFAP levels and executive functions, both within the whole cohort (r=-0.37, p = 0.01, Additional file 1, Fig. S3a) and among subjects with a CSF AD profile (r=-0.39, p = 0.04, Additional file 1, Fig. S3b). Correlation between GFAP levels and processing speed did not reach significance within the whole cohort (r=-0.27, p = 0.06, Fig. 4e) or among those with a CSF non-AD profile (r = 0.02, p = 0.94), but it did strongly correlate among those with a CSF AD profile (r=-0.68, p < 0.001, Fig. 4f). A similar albeit weaker pattern was found between CSF S100B levels and processing speed (within whole cohort: r=-0.20, p = 0.16, Fig. 4g; CSF AD profile: r=-0.45, p = 0.02; CSF non-AD profile: r = 0.03, p = 0.89, Fig. 4h). The corresponding correlations between CSF S100B and GFAP levels with individual neuropsychological tests reflecting processing speed and executive functions are presented in Additional file 1, Fig. S4a-g.