Skip to main content
. 2023 Dec 23;28(1):15. doi: 10.1007/s00784-023-05401-8

Table 3.

DerSimonian-Laird random-effects model analysis for MBL under different follow-ups

Follow-up (months) Studies*/implants (n) MBL estimate (95% CI) (mm) SE p value Heterogeneity
2–6 9/333 0.948 (0.641, 1.256) 0.157 < 0.001 τ2 = 0.196, p < 0.001, I2 = 91.941
12–15 15/697 0.632 (0.434, 0.830) 0.101 < 0.001 τ2 = 0.130, p < 0.001, I2 = 93.299
18–24 5/183 1.021 (0.274, 1.769) 0.381 0.007 τ2 = 0.704, p < 0.001, I2 = 99.577
30–36 7/413 0.892 (0.429, 1.354) 0.236  < 0.001 τ2 = 0.354, p < 0.001, I2 = 96.225
48 2/36 2.060 (1.885, 2.235) 0.089  < 0.001 τ2 = 0.045, p = 0.048, I2 = 81.493
60 5/217 0.958 (0.705, 1.211) 0.129  < 0.001 τ2 = 0.068, p < 0.001, I2 = 89.234
70–132 6/321 1.175 (0.876, 1.473) 0.152  < 0.001 τ2 = 0.123, p < 0.001, I2 = 94.685

MBL, marginal bone loss; 95% CI, 95% confidence interval; SE, standard error

*When data on MBL for all implants in a study was not available (as a global mean value), then data on the mean value of the different sub-groups was entered. In these cases, each sub-group was considered one “study”