Stepwise multivariable linear regression with QTc interval as dependent variable. To minimize risk of collinearity, these were entered into separate models: β blocker, heart disease, and hypercholesterolemia, as well as anti-CCP positivity and RF positivity. Standardized β and unstandardized (B) coefficients are shown.
| Variable | β | B | 95% CI | p |
|---|---|---|---|---|
| QTBAZ as dependent variable, stepwise linear regression | ||||
| Model 1, R2 = 0.220 | ||||
| Age | 0.263 | 0.487 | 0.175–0.800 | 0.003‡ |
| Female sex | 0.236 | 11.719 | 3.354–20.083 | 0.006‡ |
| CRP | 0.278 | 0.300 | 0.118–0.483 | 0.001‡ |
| Model 2, R2 = 0.211 | ||||
| Age | 0.264 | 0.489 | 0.169–0.808 | 0.003‡ |
| Female sex | 0.210 | 10.441 | 1.894–18.987 | 0.017† |
| β blocker | −0.177 | −9.973 | −19.647 – −0.298 | 0.043† |
| IL-10, tertiles, per unit increase | 0.202 | 5.457 | 0.754–10.160 | 0.023† |
| QTFHS as dependent variable, stepwise linear regression | ||||
| Model 3, R2 = 0.238 | ||||
| Heart disease | 0.351 | 20.321 | 10.562–30.080 | < 0.001‡ |
| HCQ | 0.248 | 11.197 | 3.650–18.744 | 0.004‡ |
| IL-10, tertiles, per unit increase | 0.223 | 4.830 | 1.205–8.455 | 0.009‡ |
The following variables were entered into the stepwise multivariable regression analysis: models 1 and 2 (age, female sex, CRP, IL-6, TNF-α, IL-1α, IL-1β, IL-10, heart disease, β blocker, and hydroxychloroquine), and Model 3 [age, female sex, TNF-α (tertiles), IL-1β (tertiles), IL-10 (tertiles), heart disease, hypercholesterolemia, hydroxychloroquine, RF positivity, and anti-CCP positivity].
↵† p < 0.05.
↵‡ p < 0.01. QTc: corrected QT; anti-CCP: anticyclic citrullinated peptide antibodies; RF: rheumatoid factor; QTBAZ: QT corrected using the Bazett formula; CRP: C-reactive protein; IL: interleukin; QTFHS: QT corrected using the Framingham Heart Study formula; HCQ: hydroxychloroquine.