Skip to main content
. 2023 Apr 11;12(5):e230001. doi: 10.57264/cer-2023-0001

Table 3. . Comparison of functional and safety outcomes between EmboTrap, Trevo, and Solitaire after removing outlier studies, with pooled random effects estimates.

Clinical and safety outcomes Studies reporting outcomes (n) Study arms reporting outcome (n) EmboTrap Trevo Solitaire p-value (EmboTrap vs Trevo) p-value (EmboTrap vs Solitaire) p-value (Trevo vs Solitaire) I2 (EmboTrap + Trevo)
(95% CI)
p-value
I2 (EmboTrap + Solitaire)
(95% CI)
p-value
I2 (Trevo + Solitaire)
(95% CI)
p-value
Overall I2
(95% CI)
p-value
mRS 0-2 at 90 days 41 44 228/416 (54.9%)
[48.7–61.1]
1667/3160 (50.0%)
[46.5–53.5]
2018/4493 (46.2%)
[43.8–48.6]
0.157 0.008 0.085 39.0% (0.0–65.2)
0.047
52.7% (29.2–68.4)
<0.001
66.3% (52.6–76.0)
<0.001
65.2% (52.2–74.7)
<0.001
mRS at 90 days 20 21 1.63 (0.94–2.32)
[475]
2.45 (1.93–2.98)
[2598]
2.73 (2.26–3.20)
[2893]
0.118 0.018 0.441 85.7% (57.2–96.9)
<0.001
89.5% (79.2–96.0)
<0.001
91.8% (83.9–96.7)
<0.001
93.4% (87.9–96.9)
<0.001
0 . . 100 (21.1%) 497 (19.1%) 360 (12.5%) . . . . . . .
1 . . 92 (19.4%) 508 (19.6%) 479 (16.6%) . . . . . . .
2 . . 98 (20.6%) 387 (14.9%) 448 (15.5%) . . . . . . .
3 . . 48 (10.1%) 332 (12.8%) 401 (13.9%) . . . . . . .
4 . . 63 (13.3%) 346 (13.3%) 354 (12.2%) . . . . . . .
5 . . 19 (4.0%) 128 (4.9%) 177 (6.1%) . . . . . . .
6 . . 55 (11.6%) 400 (15.4%) 673 (23.3%) . . . . . . .
Mortality at 90 days 38 41 51/436 (12.2%)
[9.4–15.7]
481/3172 (15.7%)
[12.4–19.6]
679/3254 (18.5%)
[16.0–21.3]
0.159 0.023 0.260 62.6% (37.8–77.5)
<0.001
73.0% (62.3–80.7)
<0.001
71.6% (61.2–79.2)
<0.001
71.6% (61.2–79.2)
<0.001
ENT or distal emboli 24 25 24/537 (4.5%)
[2.4–8.4]
50/875 (6.4%)
[4.3–9.5]
157/1961 (7.3%)
[5.5–9.6]
0.366 0.167 0.624 40.9% (0.0–70.8)
0.077
62.0% (36.7–77.2)
<0.001
58.6% (33.0–74.5)
<0.001
59.1% (36.5–73.7)
<0.001
sICH 42 44 22/687 (3.9%)
[2.3–6.6]
68/1076 (6.7%)
[5.3–8.4]
319/4629 (7.6%)
[6.4–8.9]
0.169 0.015 0.265 9.2% (0.0–46.1)
0.348
45.0% (17.9–63.1)
0.003
33.2% (0.0–55.4)
0.028
36.6% (8.4–56.1)
0.009

Dichotomous data for individual subgroups are expressed as n/n (%) [95% CI]; % calculated represents pooled random effects estimate, not fixed percentage).

Note: Outlier and influencer analyses were only performed if there were at least 5 studies per treatment group that reported the variable of interest. As such, outlier and influencer analyses were only performed for comparisons of mRS 0-2 at 90 days, mortality at 90 days, sICH, and ENT.

Ordinal data for individual subgroups are expressed as pooled median (95% CI) [n], along with raw frequency counts and percentages for each ordinal score. All pooled estimates for dichotomous data are derived from random-effects models using the DerSimonian-Laird procedure for estimation of between-study variance [15]; 95% CIs of the pooled results were computed using the Jackson method [16]. Pooled medians and corresponding 95% CIs were derived via random effects models using methods described by McGrath et al. [17] P-values for each pairwise comparison are provided using separate meta-regression analyses, considering the intervention as a categorical moderator. P-values for the overall heterogeneity (i.e., statistical inconsistency) among the included studies are obtained from Q-tests of heterogeneity. The estimated percentage of variability in effect size estimates that is due to heterogeneity rather than sampling error is given by I2 statistics and their corresponding 95% CIs [18]. I2 values are given for each subgroup comparison and for the overall study population.

CI: Confidence interval; ENT: Embolization to new territory; FPR: First pass recanalization; mRS: Modified Rankin Scale; mTICI: Modified thrombolysis in cerebral infarction; sICH: Symptomatic intracranial hemorrhage; TICI: Thrombolysis in cerebral infarction.