| Source | Treatment | No. of patients with at least 1 serious infection/total in treatment groups(incidence %) | No. of patients with at least 1 serious infection/total in placebo groups(incidence %) | Pooled ORs* (95% CI) |
| Edwards et al, 2004 | Rituximab | 6/121 (4.9) | 1/40 (2.5) | 1.45 (0.56 to 3.73) |
| Emery et al, 2006 | 4/316 (1.2) | 2/149 (1.3) | ||
| Cohen et al, 2006 | 7/308 (2.3) | 3/209 (1.4) | ||
| Total | 17/745 (2.3) | 6/398 (1.5) | ||
| Moreland et al, 2002 | Abatacept | 1/90 (1.11) | 0/32 (0) | 1.35 (0.78 to 2.32) |
| Kremer et al, 2005 | 1/220 (0.45) | 2/119 (1.7) | ||
| Genovese et al, 2005 | 6/258 (2.3) | 3/133 (2.2) | ||
| Kremer et al, 2006 | 13/433 (3.0) | 5/219 (2.3) | ||
| Weinblatt et al, 2006 | 28/959 (2.9) | 8/482 (1.6) | ||
| Total | 49/1960 (2.5) | 18/985 (1.8) | ||
| Bresnihan et al, 1998 | Anakinra | 5/351 (1.42) | 1/121 (0.82) | 2.75 (0.90 to 8.35) |
| Cohen et al, 2002 | 0/345 (0) | 0/74 (0) | ||
| Cohen et al, 2004 | 2/250 (0.80) | 2/251 (0.8) | ||
| Schiff et al, 2004 | 23/1116 (2.0) | 1/283 (0.3) | ||
| Total | 30/2062 (1.4) | 4/729 (0.5) |
*Using the Mantel–Haenszel method to calculate a pooled ORs with fixed effect.
χ2 (test OR differs from 1): for Rituximab: χ2 = 0.29, p = 0.6, for Abatacept: χ2 = 0.94, p = 0.3, for Anakinra: χ2 = 2.5, p = 0.1.