Table 5.

Meta-regression results on the separate influence of participant-related characteristics on differences in VO_2peak between the arm crank ergometer (ACE) and the wheelchair ergometer mode (WERG).

	Coefficient [95% CI]	p-value	Constant [95% CI]	p-value	Tau2	$I_{\text{res}}^2$ (%)	$R_{\text{adj}}^2$ (%)	# of subgroups
EFFECT SIZE—DIFFERENCES IN ABSOLUTE VO2peak: ACE-WERG
Body mass (kg)	−0.009 [−0.02–0.004]	0.16	0.6 [−0.2–1.4]	0.15	0	0	0	15
% of able-bodied participants	0.0006 [−0.002–0.003]	0.60	0.007 [−0.06–0.07]	0.83	0	0	0	18
% of participants with a tetraplegia	−0.0005 [−0.002–0.001]	0.41	0.04 [−0.05–0.1]	0.39	0	0	0	18
% of participants who are wheelchair athletes	0.0005 [−0.001–0.002]	0.49	−0.02 [−0.1–0.1]	0.69	0	0	0	17
EFFECT SIZE—DIFFERENCES IN BODY-MASS NORMALIZED VO2peak: ACE-WERG
Body mass (kg)	0.07 [−0.1–0.3]	0.43	−5.0 [−18.6–8.6]	0.44	2.9	44	−11	14
% of able-bodied participants	–	–	–	–	–	–	–	–
% of participants with a tetraplegia	−0.01[−0.05−0.02]	0.34	0.6 [−1.2−2.4]	0.50	2.2	32	2	15
% of participants who are wheelchair athletes	0.002 [−0.03−0.04]	0.91	−0.1 [−2.4 −2.2]	0.90	3.1	40	−13	14

Note that there are no coefficients for the influence of % of able-bodied participants on difference in body-mass normalized VO_2peak between the ACE and the WERG mode, since only one study with able-bodied participants was included for this comparison.