3.3 Subgroup analyses in selected populations
In the subgroup
analyses, the outcomes stratifying the different dosages of the drug are
consistent with the overall analysis. There was no difference between
the groups in the resolution of NASH at low (RR 4.34; 95% CI 0.60 to
31.27; p=0.14; I²=62%; Figure S5), medium (RR 3.91; 95% CI 0.85 to
16.06; p=0.08; I²=52%; Figure S5) or high doses (RR 3.54; 95% CI 0.86
to 14.49; p=0.03; I²=44%; Figure S5). The removal of each study from
the pooled analysis did not affect the NASH resolution endpoint, except
for FALCON-1. The withdrawal of this study increased the frequency of
resolution of NASH for low (RR 12.20, 95% CI 1.89-78.71; p=0.009; I2=
12%; Figure S15), medium (RR 8.73, 95% CI 2.42-31.48; p=0.0009; I2=
0%) and high doses (RR 7.39, 95% CI 2.30-23.79; p=0.0008; I2= 0%).
The use of FGF21 in low
(MD -0.75; 95% CI -0.80 to 2.31; p=0.96; I²=0%; Figure S6), medium (MD
-0.51; 95% CI -1.70 to 0.69; p=0.48; I²=0%; Figure S6) or high doses
(MD -1.26; 95% CI -3.93 to 1.42; p=0.04; I²=70%; Figure S6) did not
contribute to weight reduction, corroborating the global analysis.
Regarding adverse events, there was also no significant difference
between groups in low (RR 1.17; 95% CI 1.00 to 1.37; p=0.04; I²=60%;
Figure S7), medium (RR 1.09; 95% CI 1.00 to 1.19; p=0.24; I²=25%;
Supplementary Figure S8) or high doses (RR 1.07; 95% CI 1.00 to 1.15;
p=0; I²=80%; Supplementary Figure S9). The subanalysis results for
serious adverse events revealed no significant association between the
utilization of FGF21 analogs at low (RR 0.93; 95% CI 0.33 to 2.26;
p=0.20; I²=35%; Figure S10), medium (RR 1.10; 95% CI 0.58 to 2.08;
p=0.53; I²=0%; Figure S11) or high doses (RR 1.49; 95% CI 0.64 to
3.47; p=0.19; I²=33%; Figure S12).