Table 4.
Results of principal component analysis.
| Variables | PC1 | PC2 | PC3 | PC4 |
|---|---|---|---|---|
| Eigenvalue | 5.19 | 1.82 | 1.47 | 1.01 |
| Proportion of variance | 0.52 | 0.18 | 0.15 | 0.10 |
| Cumulative variance | 0.52 | 0.70 | 0.85 | 0.95 |
| Loadings | ||||
| SIDe | 1.00 | |||
| HCO3− | 0.99 | |||
| tCO2 | 0.99 | |||
| BE | 0.98 | |||
| corAG | −0.80 | 0.45 | ||
| pH | 0.71 | −0.66 | ||
| SIDa | 0.92 | |||
| Cl | −0.87 | |||
| pCO2 | 0.34 | 0.94 | ||
| ATot | 0.99 | |||
When the coefficient equaled to 0.3 or more, factor loading was considered significant. Factor loadings are presented here, after sorting in small orders from large numbers. As strong ion gap (SIG) was calculated by subtracting SIDe from SIDa, SIG was removed from the principal component analysis. When SIG was input with other variables, the inter-item correlation matrix failed to be positive definite. Abbreviations: SIDe, effective strong ion difference; HCO3−, bicarbonate concentration; tCO2, total concentration of carbon dioxide; BE, base excess; corAG, corrected anion gap; pH, power of hydrogen; SIDa, apparent strong ion difference; Cl, chloride; pCO2, partial pressure of carbon dioxide; ATot, total concentration of serum weak acids.