Table 2. Analysis of baseline total WBC count and WBC composition (% lymphocyte; % neutrophil) exposures and longitudinal change in CES-D score (sex-stratified), mixed-effects linear regression analysis, HANDLS study, 2004–2013.
Men: Model 2a |
Women: Model 3a |
|||
---|---|---|---|---|
γ±SEE | P-value | γ±SEE | P-value | |
N =877 | N '=1488 | N =1134 | N '=2008 | |
Total WBC count (TWBCC) | ||||
Fixed effect | ||||
Intercept (γ00 for π0i) | +19.50±2.29 | <0.001 | +20.18±2.12 | <0.001 |
Time (γ10 for π1i) | +0.50±0.52 | 0.34 | −0.13±0.49 | 0.79 |
Agebase | −0.08±0.04 | 0.031 | −0.07±0.04 | 0.034 |
Agebase × time | +0.02±0.01 | 0.08 | +0.013±0.008 | 0.13 |
TWBCC (γ011a for π0i) | ||||
<10th vs 10th−90th | +2.51±1.05 | 0.016 | −0.08±1.09 | 0.94 |
>90th vs 10th−90th | −0.47±1.23 | 0.70 | −1.13±1.16 | 0.33 |
TWBCC × time (γ111a for π1i) | ||||
<10th vs 10th−90th | −0.44±0.23 | 0.06 | +0.11±0.24 | 0.65 |
>90th vs 10th−90th | +0.40±0.29 | 0.17 | +0.81±0.28 | 0.003 |
Random effects | ||||
Level 1 residuals (Rij) | +6.07±0.27 | <0.001 | +7.60±0.18 | <0.001 |
Level 2 residuals | ||||
Intercept (ξ0i) | +7.32±0.27 | <0.001 | +8.05±0.27 | <0.001 |
Linear slope (ξ1i) | +0.45±0.26 | NS | +0.00±0.06 | NS |
Percent neutrophil (PN) | ||||
Fixed effect | ||||
Intercept (γ00 for π0i) | +20.3±2.29 | <0.001 | +19.77±2.11 | <0.001 |
Time (γ10 for π1i) | +0.39±0.52 | 0.027 | −0.13±0.49 | 0.79 |
Agebase | −0.08±0.04 | 0.027 | −0.08±0.04 | 0.035 |
Agebase × time | +0.02±0.01 | 0.09 | +0.012±0.009 | 0.15 |
PN (γ012a for π0i) | ||||
<10th vs 10th−90th | −0.10±1.03 | 0.93 | +2.70±1.26 | 0.031b |
>90th vs 10th−90th | −0.59±1.16 | 0.61 | +1.60±1.14 | 0.16 |
PN × time (γ112a for π1i) | ||||
<10th vs 10th−90th | −0.10±0.24 | 0.70 | +0.07±0.28 | 0.79 |
>90th vs 10th−90th | −0.11±0.26 | 0.69 | +0.05±0.28 | 0.86 |
Random effects | ||||
Level 1 residuals (Rij) | +6.11±0.27 | <0.001 | +7.63±0.18 | <0.001 |
Level 2 residuals | ||||
Intercept (ξ0i) | +7.33±0.27 | <0.001 | +8.00±0.27 | <0.001 |
Linear slope (ξ1i) | +0.42±0.27 | NS | +0.00±0.00 | NS |
Percent lymphocytes (PL) | ||||
Fixed effect | ||||
Intercept (γ00 for π0i) | +20.18±2.29 | <0.001 | +19.72±2.12 | <0.001 |
Time (γ10 for π1i) | +0.38±0.52 | 0.46 | −0.09±0.49 | 0.86 |
Agebase | −0.08±0.04 | 0.025 | −0.079±0.036 | 0.030 |
Agebase × time | +0.016±0.009 | 0.08 | +0.012±0.009 | 0.15 |
PL (γ013a for π0i) | ||||
<10th vs 10th−90th | +0.17±1.04 | 0.87 | +1.95±1231 | 0.11 |
>90th vs 10th−90th | +0.74±1.06 | 0.69 | +1.90±1.18 | 0.11 |
PL × time (γ113a for π1i) | ||||
<10th vs 10th−90th | −0.16±0.25 | 0.53 | −0.26±0.29 | 0.38 |
>90th vs 10th−90th | −0.11±0.25 | 0.65 | +0.001±0.27 | 0.99 |
Random effects | ||||
Level 1 residuals (Rij) | +6.11±0.27 | <0.001 | +7.63±0.18 | <0.001 |
Level 2 residuals | ||||
Intercept (ξ0i) | +7.33±0.27 | <0.001 | +8.01±0.28 | <0.001 |
Linear slope (ξ1i) | +0.43±0.27 | NS | +0.001±0.001 | NS |
Abbreviations: AA, arachidonic acid; ALA, α-linolenic acid; CES-D, Center for Epidemiologic Studies-Depression scale; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; HANDLS, healthy aging in neighborhoods of diversity across the lifespan; HDL-C, high-density lipoprotein-cholesterol; HS, high school; LA, linoleic acid; LDL-C, low-density lipoprotein-cholesterol; n-3, omega-3; n-6, omega-6; NS, not significant; PUFA, polyunsaturated fatty acids; TC, total cholesterol; WBC, white blood cell.
Models were further adjusted for other covariates (main effects and interaction with time). See the 'Materials and methods' section for more details on covariate coding and model specifications. Time at baseline visit was set to zero. Baseline age was centered at 50 years (y), total energy intake at 2000 kcal per day, total carotenoid intake at 3 mg per 1000 kcal per day, vitamin C intake at 30 mg per 1000 kcal per day, vitamin A intake at 300 RE per 1000 kcal per day, vitamin E at 3 mg per 1000 kcal per day, vitamin B-6 at 0.8 mg per 1000 kcal per day, vitamin B-12 at 3 μg per 1000 kcal per day, folate at 170 μg per 1000 kcal per day, n-3 PUFA:n-6 PUFA at 0.11. Healthy Eating Index-2010 was centered at 42.
In a separate model with interaction of WBC exposures by sex by time, including all other terms in the current model, P<0.10 for null hypothesis that this interaction term is =0.
n represents the number of participants in the analysis; n' represents the total number of visits included in the analysis. Findings that were significant at a type I error of 0.05 are in bold. Subscript 'a' refers to the exposure level whereby 1: <10th vs 10th−90th and 2: >90th vs 10th−90th. Italic entries are when P<0.10 but >0.05, so marginally significant results prior to correction for multiple testing.