study due to the constraints of their medical management
(Figure 1). All participants permitted their length of stay to be
recorded.
The baseline characteristics of the groups were similar. These
data are presented in Table 1[12TD$DIF] and the first two data columns of
Table 2, Table 3 and Table 4. Individual participant data are
presented in Table 5 on the eAddenda.
Effect of interventions
Primary outcome
Functional exercise capacity, as measured by the Glittre
Activities of Daily Living test, improved by a mean of 52 seconds
(SD 40) in the experimental group, whereas the control group
improved by a mean of 12 seconds (SD 26). This was a statistically
significant difference (MD 39 seconds, 95% CI 20 to 59), as shown in
Table 2. When calculated as a standardised effect size, the
between-group difference of 39 seconds equated to a Cohen’s d
of 1.19.
Secondary outcomes
The improvement in functional capacity, as evaluated by the
distance walked on the ISWT, was significantly greater in the
experimental group than in the control group (MD 130 m, 95% CI
77 to 182). When calculated as a standardised effect size, the
between-group difference equated to a Cohen’s d of 1.39. No
significant between-group differences occurred in the variables
monitored during the ISWT: heart rate, percentage of predicted
maximal heart rate, SpO2 and Borg dyspnoea and fatigue scales
(data not shown). The SpO2 during the aerobic training did not
show any desaturation < 84%, as outlined in the Methods section.
All muscles that were analysed demonstrated mean improvement
in strength in the experimental group and mean deterioration
in the control group, as presented in Table 2. The betweengroup
difference in change in strength was statistically significant
for all four muscle groups. The between-group differences in change
in strengthwere: 3.5 kgf for biceps brachii (Cohens’ d = 0.76); 2.2 kgf