Table 4.
Results of multivariate logistic regression analysis of factors associated with death in invasive GAS infectiona
| Variable | Odds ratio (95% CI) | P value |
|---|---|---|
| Age <50 yr | Reference | |
| Age ≥50 yr | 2.9 (1.5-5.6) | <0.01 |
| Toxic shock syndrome | 9.9 (5.6-17.5) | <0.001 |
| Bacteremia without identified focus | 2.5 (1.4-4.7) | <0.01 |
| Chemotherapy | 3.9 (0.9-16.2) | 0.06 |
| Diabetes | 1.8 (0.9-3.6) | 0.10 |
| Liver disease | 3.7 (1.3-10.5) | 0.02 |
| emm1 type | 2.2 (1.2-3.8) | <0.01 |
a
Data are for 600 cases. Clinical presentations, predisposing factors, emm types, and patient characteristics were entered in the model if associated with an increased or a decreased frequency of death on univariate analysis (P < 0.20) and retained in the model if associated with death (P < 0.10). Finally, factors were considered to be independently associated with death if associated with a P value of 0.05 or less (Wald test). Patients with missing data (n = 64) were excluded from the analysis. Two-way interactions were evaluated.