Skip to main content
. 2021 Aug 21;25(4):615–626. doi: 10.1038/s41391-021-00442-0

Table 4.

Overall and subgroup effects of low muscle mass vs. high muscle mass on overall survival in prostate cancer patients.

Analysis No. of comparisons Sample size Main effect Subgroup differences
HR (95% CI) I2 p value χ2 p value
Multivariate analysis
 Overall effect 9 3275 1.50 (1.10–2.05) 88% 0.009
 Overall effect without outlier 8 3093 1.63 (1.27–2.08) 58% <0.001
Population subgroups
  Advanced diseasea 5 723 1.43 (1.07–1.92) 34% 0.020 1.8 0.184
  Localised disease 3 2370 1.91 (1.40–2.62) 52% <0.001
Previously defined cut-off values
  Yesa 5 1751 1.50 (1.01–2.22) 56% 0.036 0.4 0.532
  No 3 1342 1.77 (1.27–2.48) 67% <0.001
Multivariate models controlling for BMI
  Yesa 4 1574 1.48 (0.98–2.26) 77% 0.060 0.5 0.502
  No 4 1519 1.77 (1.32–2.36) 17% <0.001
Univariate analysis
 Overall effect 13 2638 1.40 (1.13–1.72) 64% 0.002
 Overall effect without outlier 12 1618 1.31 (1.09–1.58) 39% 0.004
Population subgroups
  Advanced disease 11 1565 1.32 (1.09–1.61) 44% 0.005 0.7 0.413
  Localised disease 2 1073 1.74 (0.93–3.25) 62% 0.083
Previously defined cut-off values
  Yes 6 832 1.32 (0.80–2.17) 67% 0.271 0.0 0.974
  Noa 6 1099 1.33 (1.15–1.54) 0% <0.001

BMI body mass index, HR hazard ratio, I2 indicator of heterogeneity (%).

aAdjustment after sensitivity analysis omitting one study at a time.