Table 4.
Cytokine secretion as predictors of self-reported cold or flu incidence and severity.
| LPS-induced IL-6 secretion from PBMCs |
Anti-CD3-induced IFN-γ secretion from T cells |
|||||||
|---|---|---|---|---|---|---|---|---|
| β | Odds ratio (95% CI) | Variables in the model | p value | β | Odds ratio (95% CI) | Variables in the model | p value | |
| Incidence of cold or flu episode in the past montha,b | ||||||||
| Model 1 | −0.1471 | 0.863 (0.749, 0.994) | IL-6 secretion | 0.0414 | 0.0141 | 1.014 (0.997, 1.032) | IFN-γ secretion | 0.1019 |
| Model 2 | −0.2392 | 0.787 (0.632, 0.980) | Model 1 + total calories, vitamin C, iron, zinc | 0.0325 | 0.0556 | 1.057 (0.986, 1.133) | Model 1 + age, BMI, vitamin D, iron, zinc, n-3 PUFA, alcohol, caffeine | 0.1158 |
| β | 95% CI | Variables in the model | p value | β | 95% CI | Variables in the model | p value | |
| Total number of days with cold or flu symptoms in the past monthc | ||||||||
| Model 1 | 0.0252 | −0.042, 0.092 | IL-6 secretion | 0.4609 | −0.0138 | −0.022, −0.006 | IFN-γ secretion | 0.0005 |
| Model 2 | 0.0801 | −0.018, 0.178 | Model 1 + total calories, iron, zinc | 0.1105 | −0.0153 | −0.021, −0.009 | Model 1 + vitamin C, zinc, alcohol, caffeine | <0.0001 |
aLogistic regression model is fitted to determine the relationship between IL-6 or IFN-γ secretion and incidence of cold or flu in the past month.
bNo cold or flu episode in the past month = 0, presence of a cold or flu episode in the past month = 1. Probability modeled is no cold or flu episode in the past month.
cPoisson regression model is fitted to determine the relationship between IL-6 or IFN-γ secretion and total number of days with cold or flu symptoms in the past month.