Table 3.

Mediation effects for knee outcomes: association of GLP-1RA therapies with the incidence of knee surgery

Exposure: GLP-1RA	Mediator: weight change from baseline Incidence of knee surgery* (95% CI)	P value	Mediator: HbA1c change from baseline Incidence of knee surgery† (95% CI)	P value
Controlled direct effect	−0.027 (−0.050, 0.008)‡	0.095	−0.041 (−0.060, −0.012)‡	0.014
Indirect effect	−0.015 (−0.027, −0.005)	0.004	0.000 (−0.001, 0.002)	0.78
Total effect	−0.042 (−0.061, −0.016)	0.006	−0.040 (−0.060, −0.012)	0.015
Proportion mediated	32.1% (11.0%, 142.8%)	0.010	Proportion too small to calculate—not a mediator	0.78

*The model adjusted for age, sex, baseline BMI, baseline Kellgren-Lawrence grade and baseline WOMAC total score.

†The model adjusted for age, sex, baseline HbA1c, baseline BMI, baseline Kellgren-Lawrence grade and baseline WOMAC total score.

‡Values are unstandardised regression coefficients representing incidence of knee surgery.

BMI, body mass index; GLP-1RA, glucagon-like peptide-1 receptor agonist; HbA1c, hemoglobin A1c; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.