Table 4.
Multi-analyte regression models1
| (a) Depression-Total (BDI-II) | |||
|---|---|---|---|
| Model fit | F(1, 77) = 4.0145; | P = 0.0487; | R2 = 0.0502 |
| Variable | b | t | P |
| Intercept | 1.8205 | 0.80 | 0.4284 |
| HCV status | 2.8974 | 2.00 | 0.0487 |
| Model fit | F(8, 77) = 4.787; | P < 0.001; | R2 = 0.3567 |
| Variable | b | t | P |
| Intercept | 17.3226 | 4.20 | <0.0001 |
| HCV status | 5.1596 | 3.23 | 0.0019 |
| A2Macro | −12.1283 | −2.02 | 0.0475 |
| BDNF | −1.4825 | −2.89 | 0.0052 |
| Eotaxin1 | −0.0165 | −3.15 | 0.0024 |
| IL23 | 4.7228 | 2.96 | 0.0042 |
| RANTES | 0.4548 | 1.99 | 0.0509 |
| TNFα | 2.6492 | 3.32 | 0.0015 |
| TNFR2 | −3.1421 | −4.29 | <0.0001 |
| (b) Depression-Cognitive Affective Factor (BDI-II) | |||
| Model fit | F(1, 77) = 1.364; | P = 0.2466; | R2 = 0.0176 |
| Variable | b | t | P |
| Intercept | 0.0862 | 0.80 | 0.4241 |
| HCV status | 0.0793 | 1.17 | 0.2466 |
| Model fit | F(11, 77) = 4.0268; | P < 0.0002; | R2 = 0.4016 |
| Variable | b | t | P |
| Intercept | 0.2661 | 1.35 | 0.1831 |
| HCV status | 0.1556 | 2.12 | 0.0375 |
| AAT | 0.4588 | 2.52 | 0.0143 |
| BDNF | −0.0731 | −3.11 | 0.0028 |
| Eotaxin1 | −0.0005 | −2.03 | 0.0465 |
| IL15 | −0.1373 | −2.05 | 0.0441 |
| IL18 | −0.0011 | −1.67 | 0.0997 |
| IL23 | 0.2184 | 3.00 | 0.0039 |
| RANTES | 0.0221 | 2.10 | 0.0394 |
| TNFα | 0.1198 | 3.22 | 0.0020 |
| TNFR2 | −0.2101 | −4.96 | <0.0001 |
| vWF | 0.0088 | 1.93 | 0.0579 |
| (c) Depression–Somatic Factor (BDI-II) | |||
| Model fit | F(1/76) = 6.1293, | P = 0.0156, | R2 = 0.0756 |
| Variable | b | t | P |
| Intercept | 0.1001 | 0.75 | 0.4559 |
| HCV status | 0.2082 | 2.48 | 0.0156 |
| Model fit | F(9/76) = 3.2644 | P = 0.0024, | R2 = 0.3048 |
| Variable | b | t | P |
| Intercept | 1.0215 | 3.25 | 0.0018 |
| HCV status | 0.3342 | 3.49 | 0.0009 |
| A2Macro | −0.9127 | −2.39 | 0.0196 |
| C3 | 1.1813 | 1.75 | 0.0841 |
| Fibrinogen | −0.1525 | −1.88 | 0.0649 |
| IL23 | 0.2075 | 2.07 | 0.0426 |
| IL8 | 0.0146 | 1.95 | 0.0558 |
| MCP1 | −0.0031 | −2.56 | 0.0128 |
| MIP1β | −0.0020 | −2.57 | 0.0123 |
| MMP3 | −0.0538 | −2.14 | 0.0363 |
| (d) Anxiety (GADI) | |||
| Model fit | F(1, 76) = 5.4028; | P < 0.0228; | R2 = 0.0672 |
| Variable | b | t | P |
| Intercept | 2.0202 | 0.65 | 0.5184 |
| HCV status | 4.5951 | 2.32 | 0.0228 |
| Model fit | F(5, 76) = 4.7444; | P < 0.009; | R2 = 0.2504 |
| Variable | b | t | P |
| Intercept | 0.6811 | 0.17 | 0.8628 |
| HCV status | 4.9587 | 2.13 | 0.0370 |
| MIP1α | 0.1114 | 1.68 | 0.0966 |
| SCF | 0.0487 | 1.92 | 0.0589 |
| TNFα | 2.3309 | 1.98 | 0.0516 |
| TNFR2 | −3.1440 | −3.20 | 0.0020 |
| (e) Fatigue (FSS) | |||
| Model fit | F(1, 75) = 9.0997; | P < 0.035; | R2 = 0.1095 |
| Variable | b | t | P |
| Intercept | 1.6061 | 3.06 | 0.0030 |
| HCV status | 1.0000 | 3.02 | 0.0035 |
| Model fit | F(8, 75) = 4.3840; | P < 0.0003; | R2 = 0.3436 |
| Variable | b | t | P |
| Intercept | 0.7798 | 0.78 | 0.4359 |
| HCV status | 0.5099 | 1.46 | 0.1485 |
| AAT | 2.1043 | 2.17 | 0.0335 |
| BDNF | −0.2359 | −2.10 | 0.0396 |
| FactorVII | 0.0019 | 1.93 | 0.0583 |
| IL7 | 0.1587 | 3.17 | 0.0023 |
| RANTES | 0.1255 | 2.38 | 0.0200 |
| VDBP | −0.0042 | −1.99 | 0.0511 |
| VEGF | −0.0076 | −2.11 | 0.0385 |
| (f) Pain Severity (BPI-PS) | |||
| Model fit | F(1, 76) = 1.4521; | P < 0.2320; | R2 = 0.0190 |
| Variable | b | t | P |
| Intercept | 1.2919 | 1.55 | 0.1253 |
| HCV status | 0.6376 | 1.21 | 0.2320 |
| Model fit | F(5, 76) = 3.2423; | P < 0.0108; | R2 = 0.1859 |
| Variable | b | t | P |
| Intercept | 0.9130 | 0.81 | 0.4195 |
| HCV status | 0.7797 | 1.47 | 0.1468 |
| IL10 | −0.2597 | −1.82 | 0.0722 |
| IL5 | 0.0642 | 2.26 | 0.0267 |
| MIP1β | −0.0111 | −2.28 | 0.0255 |
| SCF | 0.0148 | 2.18 | 0.0325 |
| (g) Pain Interference (BPI-PI) | |||
| Model fit | F(1, 75) = 4.1609; | P < 0.0449; | R = 0.0532 |
| Variable | b | t | P |
| Intercept | 0.2677 | 0.37 | 0.7155 |
| HCV status | 0.9513 | 2.04 | 0.0449 |
| Model fit | F(5, 75) = 3.5498; | P < 0.0064; | R2 = 0.2023 |
| Variable | b | t | P |
| Intercept | 0.6042 | 0.71 | 0.4774 |
| HCV status | 1.0427 | 2.16 | 0.0338 |
| CRP | 0.2380 | 1.80 | 0.0755 |
| IL10 | −0.2695 | −2.04 | 0.0456 |
| MMP3 | −0.3068 | −2.19 | 0.0316 |
| TNFα | 0.5051 | 1.93 | 0.0574 |
BDI-II, Beck Depression Inventory-II; BPI-PI, Brief Pain Inventory-Pain Interference; BPI-PS, Brief Pain Inventory-Pain Severity; BDNF, brain-derived neurotrophic factor, CRP, C-reactive protein; DV, dependent variable; FSS, Fatigue Severity Scale; GADI, Generalized Anxiety Disorder Inventory; HCV+, adults with chronic hepatitis C virus infection; HCV−, Adults with no history of infection with the hepatitis C virus; MIP, macrophage inflammatory protein; RANTES, Regulated upon Activation, Normal T-cell Expressed, and Secreted; TNF, tumor necrosis factor; TNFR, Tumor Necrosis Factor Receptor 2; vWF, von Willebrand factor.
Regression models were developed in order to find which combination of plasma immune factors was significantly predictive of neuropsychiatric symptom severity on each of the seven neuropsychiatric variables within the total sample. Models were constructed with a backward selection linear regression of 33 immune factors. The backward selection started with the 33 immune factors and systematically eliminated from the model variables that were not significant, retaining only those with P-values (P ≤ 0.10). HCV Status was not allowed to be eliminated. Based on the final solution of the backward regression, a two-step model for each neuropsychiatric variable (DV) was constructed and these are presented above; fit parameters are presented as well as the unstandardized regression weights (b), t values and P-values for each immune factor. In these models, the first step consisted of regressing the DV onto HCV status (coded 0 for the HCV− control HCV Status, and 1 for the HCV+ HCV Status). In the second step, the significant immune factors from the backward selection were entered simultaneously with HCV status to create the final model. See Table 1 for immune factor abbreviations.