Meta-analyses of plasma exchange in GPA and MPA
| Study ID | Intervention | Control | Outcome | Included studies | No. of participants | Summary estimate | Effects size | Heterogeneity |
| Mortality at 1 year | ||||||||
| Bellos et al27 | PLEX | No PLEX | Mortality (12 months) | 3 RCT, 4 NRIS | 427 | RR (95% CI) | 0.73 (0.40 to 1.34) | I2=10% |
| Walters et al15 | PLEX | No PLEX | Death (1 year) | 5 RCT | 267 | RR (95% CI) | 1.04 (0.57 to 1.92) | I2=19% |
| Walsh et al96 | PLEX | No PLEX | Mortality (12 months) | 6 RCT | 857 | RR (95% CI) | 0.90 (0.64 to 1.27) | I2=0% |
| Yamada et al100 | PLEX | No PLEX | Mortality (1 year) | 3 RCT | 123 | RR (95% CI) | 0.71 (0.27 to 1.86) | I2=7% |
| Zhu et al101 | PLEX | No PLEX | Mortality (1 year) | 4 RCT, 5 NRIS | 1272 | OR (95% CI) | 0.83 (0.60 to 1.14) | I2=9% |
| Mortality (any timepoint) | ||||||||
| Bellos et al27 | PLEX | No PLEX | Overall mortality (time-to-event) | 2 RCT, 2 NRIS | 965 | HR (95% CI) | 0.96 (0.72 to 1.29) | I2=0% |
| Springer et al16 | PLEX | No PLEX | Mortality | 6 RCT | 251 | RR (95% CI) | 1.15 (0.78 to 1.70) | I2=0% |
| Walters et al15 | PLEX | No PLEX | Death at any timepoint | 6 RCT | 957 | RR (95% CI) | 0.96 (0.72 to 1.29) | I2=0% |
| Walsh et al96 | PLEX | No PLEX | Mortality (long-term FU, median 3 years) | 8 RCT | 1028 | RR (95% CI) | 0.93 (0.73 to 1.19) | I2=0% |
| Yamada et al100 | PLEX | No PLEX | Mortality (whole FU) | 4 RCT | 827 | RR (95% CI) | 0.93 (0.70 to 1.24) | I2=0% |
| Zhu et al101 | PLEX | No PLEX | Mortality (end of FU) | 1 RCT, 3 NRIS | 512 | OR (95% CI) | 1.18 (0.78 to 1.79) | I2=0% |
| ESKD at 1 year | ||||||||
| Bellos et al27 | PLEX | No PLEX | ESKD (12 months) | 3 RCT, 5 NRIS | 489 | RR (95% CI) | 1.32 (0.53 to 3.25) | I2=66% |
| Walters et al15 | PLEX | No PLEX | Dialysis at 1 year | 6 RCT | 235 | RR (95% CI) | 0.45 (0.29 to 0.72) | I2=0% |
| Walsh et al96 | PLEX | No PLEX | ESKD (12 months) | 7 RCT | 829 | RR (95% CI) | 0.62 (0.39 to 0.98) | I2=15% |
| Yamada et al100 | PLEX | No PLEX | Renal failure (death-censored, 1 year) | 2 RCT | 84 | RR (95% CI) | 0.40 (0.11 to 1.50) | I2=9% |
| Zhu et al101 | PLEX | No PLEX | ESKD (12 months) | 4 RCT, 4 NRIS | 1168 | OR (95% CI) | 0.90 (0.40 to 2.03) | I2=66% |
| ESKD (any timepoint) | ||||||||
| Bellos et al27 | PLEX | No PLEX | Overall ESKD (time-to-event) | 4 RCT, 2 NRIS | 1007 | HR (95% CI) | 0.71 (0.55 to 0.92) | I2=0% |
| Springer et al16 | PLEX | No PLEX | ESKD | 2 RCT | 841 | HR (95% CI) | 0.67 (0.40 to 1.11)* | I2=49% |
| Springer et al16 | PLEX | No PLEX | ESKD | 2 RCT | 841 | HR (95% CI) | 0.72 (0.53 to 0.98)† | I2=49% |
| Springer et al16 | PLEX | No PLEX | ESKD | 6 RCT | 251 | RR (95% CI) | 0.61 (0.42 to 0.90) | I2=20% |
| Walsh et al96 | PLEX | No PLEX | ESKD (long-term FU, median 3 years) | 7 RCT | 996 | RR (95% CI) | 0.79 (0.58 to 1.08) | I2=14% |
| Yamada et al100 | PLEX | No PLEX | Renal failure (death-censored, overall FU) | 4 RCT | 827 | RR (95% CI) | 0.85 (0.57 to 1.28) | I2=27% |
| Zhu et al101 | PLEX | No PLEX | ESKD (end of FU) | 1 RCT, 3 NRIS | 462 | OR (95% CI) | 1.15 (0.56 to 2.36) | I2=57% |
| ESKD or death (composite) | ||||||||
| Yamada et al100 | PLEX | No PLEX | ESKD or death (overall FU) | 3 RCT | 795 | RR (95% CI) | 0.97 (0.80 to 1.18) | I2=0% |
| Yamada et al100 | PLEX | No PLEX | ESKD or death (1 year) | 2 RCT | 756 | RR (95% CI) | 0.81 (0.62 to 1.06) | I2=0% |
| Infections | ||||||||
| Bellos et al27 | PLEX | No PLEX | Overall incidence of infections | 2 RCT, 5 NRIS | 498 | RR (95% CI) | 1.05 (0.63 to 1.76) | I2=53% |
| Springer et al16 | PLEX | No PLEX | Severe infections | 4 RCT | 887 | RR (95% CI) | 1.19 (0.99 to 1.43) | I2=0% |
| Walters et al15 | PLEX | No PLEX | Serious infections | 5 RCT | 956 | RR (95% CI) | 1.26 (1.03 to 1.54) | I2=0% |
| Walsh et al96 | PLEX | No PLEX | Serious infection (12 months) | 4 RCT | 647 | RR (95% CI) | 1.27 (1.08 to 1.49) | I2=0% |
| Walsh et al96 | PLEX | No PLEX | Serious infection (long-term FU, median 3 years) | 6 RCT | 957 | RR (95% CI) | 1.13 (1.03 to 1.24) | I2=0% |
| Yamada et al100 | PLEX | No PLEX | Infections | 2 RCT | 756 | RR (95% CI) | 1.20 (0.98 to 1.46) | I2=0% |
| Zhu et al101 | PLEX | No PLEX | Serious infections | 2 RCT, 4 NRIS | 1011 | OR (95% CI) | 1.25 (0.77 to 2.01) | I2=36% |
| Adverse events | ||||||||
| Bellos et al27 | PLEX | No PLEX | SAEs | 2 RCT, 5 NRIS | 498 | RR (95% CI) | 1.04 (0.59 to 1.81) | I2=71% |
| Springer et al16 | PLEX | No PLEX | SAEs | 3 RCT | 183 | RR (95% CI) | 1.04 (0.74 to 1.46) | I2=0% |
| Yamada et al100 | PLEX | No PLEX | Adverse events (whole FU) | 2 RCT | 756 | RR (95% CI) | 1.10 (0.73 to 1.68) | I2=37% |
| Zhu et al101 | PLEX | No PLEX | SAEs | 2 RCT, 2 NRIS | 1020 | OR (95% CI) | 1.07 (0.57 to 2.03) | I2=67% |
| Remission/Relapse | ||||||||
| Springer et al16 | PLEX | No PLEX | Remission | 2 RCT | 736 | RR (95% CI) | 1.34 (0.64 to 2.8) | I2=72% |
| Yamada et al100 | PLEX | No PLEX | Relapse (overall FU) | 3 RCT | 775 | RR (95% CI) | 0.81 (0.55 to 1.18) | I2=0% |
I2 refers to the proportion of overall variability that rises from between-study heterogeneity.
*Random effects model.
†Fixed effects model.
ESKD, end-stage kidney disease; FU, follow-up; GPA, granulomatosis with polyangiitis; MPA, microscopic polyangiitis; NRIS, non-randomised intervention study; PLEX, plasma exchange; RCT, randomised controlled trial; RR, risk ratio; SAE, serious adverse event.