Table 3.
Prognostic markers of survival
| Prognostic Marker (in quartiles or binary) | Unadjusted Analysis* | Adjusted Analysis* | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Estimated median survival (95% CI) in months 1, by marker quartile | Covariates included: Core clinical predictors 2 | Covariate included: ENCALS predictor score 3 | |||||||
| Q1 / No | Q2 | Q3 | Q4 / Yes | p-value | HR (95% CI) [Q4 vs Q1] 4 | p-value | HR (95% CI) [Q4 vs Q1] 5 | p-value | |
| Sex, male | 25 (22-37) | -- | -- | 31 (26-46) | 0.62 | 1.13 (0.7-1.8) | 0.61 | 0.98 (0.6-1.5) | 0.91 |
| Age at onset, years | 47 (31-48) | 25 (21-37) | 24 (17-37) | 22 (16-34) | 0.023 | 1.33 (0.3-6.2) | 0.28 | 1.34 (0.7-2.6) | 0.59 |
| Bulbar symptoms at onset | 33 (30-46) | -- | -- | 22 (17-27) | 0.039 | N/A | 1.31 (0.8-2.0) | 0.25 | |
| Diagnostic delay | 30 (22-37) | 27 (20-48) | 27 (22-ne) | 56 (21-ne) | 0.096 | 0.65 (0.1 - 3.1) | 0.73 | 1.66 (0.7-3.7) | 0.43 |
| Baseline ΔFRS | 47 (34-ne) | 35 (22-56) | 30 (22-43) | 17 (14-27) | <0.0001 13 | 3.64 (1.3-10.5) | 0.11 | 1.62 (0.7-3.7) | 0.39 |
| Baseline age | 46 (30-48) | 23 (19-37) | 25 (19-43) | 30 (17-34) | 0.073 | 0.46 (0.1-2.2) | 0.10 | 1.21 (0. 6-2.3) | 0.23 |
| Baseline ALSFRS-R | 22 (16-26) | 24 (21-32) | 48 (31-ne) | 43 (31-48) | 0.0010 13 | 0.55 (0.3-1.1) | 0.03 | 0.61 (0.3-1.1) | 0.04 |
| Baseline SVC %predicted | 20 (15-27) | 31 (23-47) | 34 (24-ne) | 47 (27-ne) | 0.0046 | 0.49 (0.3-0.9) | 0.10 | 0.64 (0.3-1.2) | 0.58 |
| Baseline ECAS total 6 | 31 (15-56) | 30 (22-47) | 31 (19-48) | 48 (24-ne) | 0.54 | 0.99 (0.5-2.1) | 0.78 | 0.77 (0.4-1.6) | 0.75 |
| Baseline ECAS ALS-specific 6 | 37 (18-ne) | 30 (20-47) | 23 (17-49) | 48 (26-ne) | 0.14 | 1.12 (0.5-2.6) | 0.12 | 0.88 (0.4-2.0) | 0.17 |
| Baseline ECAS ALS non-specific 6 | 25 (14-37) | 47 (26-48) | 30 (19-ne) | 31 (23-48) | 0.37 | 0.68 (0.3-1.4) | 0.73 | 0.65 (0.3-1.3) | 0.66 |
| Baseline cognitive impairment 6 | 30 (25-46) | -- | -- | 37 (13-ne) | 0.99 | 0.70 (0.3-1.5) | 0.41 | 0.84 (0.3-1.7) | 0.67 |
| Baseline behavioural impairment 6 | 34 (25-47) | -- | -- | 18 (8-ne) | 0.48 | 1.06 (0.4-2.4) | 0.90 | 1.67 (0.7-3.6) | 0.22 |
| Baseline ENCALS predictor score | 48 (39-ne) | 35 (26-ne) | 24 (20-32) | 17 (13-25) | <0.0001 13 | 4.55 (1.5 -14.7) | 0.054 | 1.61 (0.3-7.4) | 0.87 |
| Baseline serum NfL | 49 (46-ne) | 30 (23-35) | 26 (17-39) | 17 (13-22) | <0.0001 13 | 7.71 (3.7 -17.1) | <0.0001 13 | 7.34 (3.7-15.8) | <0.0001 13 |
| Baseline serum pNfH | 43 (30-ne) | 23 (17-37) | 25 (23-46) | 26 (19-47) | 0.18 | 1.74 (1.0-3.2) | 0.28 | 1.68 (1.0-3.0) | 0.36 |
| Baseline urinary p75ECD | 34 (24-46) | 31 (21-ne) | 22 (14-ne) | 30 (19-48) | 0.77 | 0.92 (0.4-1.9) | 0.94 | 0.65 (0.3-1.3) | 0.57 |
| Baseline serum uric acid | 25 (21-37) | 31 (15-46) | 30 (23-34) | 47 (25-56) | 0.25 | 0.69 (0.4-1.3) | 0.38 | 0.58 (0.3-1.1) | 0.22 |
| Baseline serum creatinine | 31 (21-46) | 24 (19-48) | 31 (22-ne) | 30 (23-47) | 0.99 | 0.87 (0.5-1.6) | 0.89 | 0.95 (0.5-1.7) | 0.97 |
| Baseline serum albumin | 25 (20-35) | 23 (19-52) | 32 (22-47) | 34 (27-47) | 0.72 | 0.95 (0.5-1.8) | 0.99 | 0.96 (0.5-1.8) | 0.96 |
| Baseline serum CRP | 30 (24-39) | 26 (9-ne) | 30 (22-47) | 0.85 | 1.07 (0.6-1.8) | 0.96 | 1.08 (0.6-1.7) | 0.95 | |
| Baseline plasma miR-181ab | 35 (30-ne) | 34 (24-52) | 26 (20-47) | 22 (17-30) | 0.021 | 1.90 (1.0-3.6) | 0.16 | 1.82 (1.0-3.3) | 0.096 |
| Baseline miR-181ab > 39,300 UMI 7 | 33 (26-48) | -- | -- | 23 (17-32) | 0.0062 | 1.55 (0.9-2.5) | 0.075 | 1.55 (1.0-2.4) | 0.050 |
| Baseline miR-181ab > 24,590 UMI 8 | 35 (30-48) | -- | -- | 23 (20-32) | 0.016 | 1.65 (1.1-2.6) | 0.030 | 1.73 (1.1-2.7) | 0.014 |
| Baseline NfL+miR181ab poor Px 9 | 37 (30-48) | -- | -- | 17 (15-21) | <0.0001 13 | 2.69 (1.6-4.4) | 0.0001 13 | 2.61 (1.7-4.1) | <0.0001 13 |
| Baseline NfL+miR181ab poor Px 10 | 47 (34-48) | -- | -- | 20 (16-22) | <0.0001 13 | 3.14 (2.0-5.1) | <0.0001 13 | 3.19 (2.0-5.0) | <0.0001 13 |
| Baseline NfL median split 11 | 47 (31-48) | -- | -- | 20 (16-25) | <0.0001 13 | 2.24 (1.4-3.6) | 0.0006 13 | 2.29 (1.5-3.6) | 0.0002 13 |
| Baseline NfL 4-level split 12 | 47 (46-52) | 24 (19-31) | 32 (14-18) | 17 (15-22) | <0.0001 13 | 4.28 (2.4-8.0) | 0.0001 13 | 4.12 (2.3-7.5) | 0.0001 13 |
Survival analysis. Q1-Q4 indicate quartiles of continuous predictors, with higher quartiles representing higher values. Yes/No in column headings captures the presence/absence of binary predictors. ne = not estimable.
Without inclusion of covariate or prognostic marker in the model, median survival (95% CI) = 30 (17-48) months.
Core clinical predictors in survival analyses include bulbar onset, diagnostic delay, ΔFRS, and baseline age.
ENCALS predictor score is derived from ΔFRS, bulbar onset, diagnostic delay, age at onset, SVC percent predicted, El Escorial definite ALS, presence of FTD, and presence of a C9orf72 repeat expansion.
These HRs compare Q4 to Q1 of each prognostic marker in a model that also adjusts for the core clinical predictors of survival. While the adjusted analyses include diagnostic delay, ΔFRS, and baseline age as linear covariates, the potential additional prognostic value of each of these predictors (see respective rows) is evaluated by contrasting top and bottom quartiles to detect possible non-linear effects.
These HRs compare Q4 to Q1 of each prognostic marker in a model that also adjusts for the ENCALS score as a linear covariate. The potential additional prognostic value of the ENCALS predictor score (see row) is evaluated by contrasting top and bottom quartiles to detect possible non-linear effects.
Among English speakers (n=171)
Threshold of 39,300 UMI in plasma as defined by Magen et al 11
Median of 24,590 UMI in plasma in the current dataset
Poor prognosis based on published optimal combination of NfL and miR-181ab, in which a poor prognostic factor is defined as either (NfL > 109.8pg/ml) or (NfL > 59.0pg/ml and miR-181ab > 39,300 UMI) 11.
Poor prognosis based on recalculated combination of NfL and miR-181ab using thresholds obtained from the current dataset; a poor prognostic factor is defined as either (NfL > 80.8 pg/mL) or (NfL > 44.8 pg/mL and mIR-181ab > 24,590 UMI).
Median serum NfL = 67.9 pg/mL
Serum NfL 4-level split is at the 33rd, 50th, and 67th percentiles (44.8 pg/mL, 67.9 pg/mL, and 80.8 pg/mL, respectively), i.e., tertiles and median, rather than quartiles, to mimic construction of the NfL+miR18ab measure 11.
p-value remains statistically significant after adjustment for multiplicity. Holm-Bonferroni adjusted p-values are reported in eTable 7.