Exercise and Protein Effects on Strength and Function With Weight Loss in Older Women

Ellen M. Evans; Chad R. Straight; Rachelle A. Reed; Alison C. Berg; David A. Rowe; Mary Ann Johnson

Disclosures

Med Sci Sports Exerc. 2021;53(1):183-191. 

In This Article

Results

Participant Characteristics

Of the 254 individuals assessed for eligibility, 81 were randomized postscreening to the three treatment groups (n = 27 each) with nine withdrawing before baseline testing and/or interventions for a total of 72 (PRO + EX = 23; PRO = 25; CON + EX = 24) (Figure 1; Table 1). Throughout the intervention, 11 withdrew or were excluded after baseline measures (n = 8) or after midpoint measures (n = 3) due to noncompliance with the protocol (lack of exercise training attendance or nutritional meetings; n = 6), non–study-related illness/injury (n = 3) or other reasons (time commitment; n = 2). Thus, a total of 61 participants completed the trial (PRO + EX = 19; PRO = 20; CON + EX = 22) with 75% and 85% retention rates from the randomization and intervention commencement, respectively. In comparison with the completers (n = 61), noncompleters (n = 11) were not different at baseline on age, body weight, relative adiposity, total energy intake, or protein intake (P > 0.05). The sample that completed the intervention (n = 61) was 92% White, and based on study design, the cohort was on average obese (BMI = 31.1 ± 5.1 kg·m−2). The completed intervention groups (three groups; n = 61) did not significantly differ on age, height, weight, BMI, or relative adiposity at baseline (P > 0.05). All primary outcomes of interest were normally distributed except the UPGO functional test, which was not normally distributed at two different time points and did not resolve with transformation. Thus, these outliers were removed (PRO + EX, n = 1; CON + EX, n = 1).

Figure 1.

Consort table summarizing recruitment, enrollment, and retention of participants.

Intervention Adherence

Energy and Macronutrient Intake. At baseline, groups did not differ in energy intake or macronutrient intake, expressed in grams per day or as percent of energy intake (all P > 0.05) (Table 2). All groups reduced their reported energy intake similarly (P > 0.05) throughout the intervention with an average reduction of ~2093 kJ (~500 kcal·d−1) or ~30%. Per the protocol, the higher protein groups, PRO + EX and PRO, reported intakes of 29.1% ± 5.6% of energy from protein, 39.6% ± 4.6% from carbohydrate, and 30.2% ± 5.3% of energy from fat. The conventional protein control diet group (CON + EX) reported consuming 19.3% ± 2.2% of energy from protein, 49.2% ± 6.5% of energy from carbohydrate, and 29.1% ± 4.8% of energy from fat. Also, per the protocol, PRO + EX and PRO had greater protein intake, regardless of whether expressed as absolute or relative to body weight or lean body mass (P < 0.001), lower carbohydrate (P < 0.05), and similar dietary fat intakes (P > 0.05) compared with CON + EX.

Exercise and Physical Activity. Per established protocol, participants randomized to the PRO + EX and CON + EX groups were individually progressed with optimal training overloads for the components of the exercise intervention being achieved by week 8 (data not shown). The two EX groups did not differ in adherence to the supervised EX program and achieved ~75% attendance. Physical activity outside the supervised intervention did not differ at baseline among the groups (P > 0.05) for either daily steps or daily minutes of MVPA. Across the intervention period, PRO + EX and CON + EX increased steps and MVPA compared with PRO, as expected based on the research design and protocol (P < 0.05; data not shown).

Weight and Body Composition

Weight. Participants who completed the intervention (n = 61) lost −9.2% ± 4.8% of initial body weight, and 42.6% met the weight loss goal of 10% of initial body weight (Table 3). There was no significant group–time interaction effect for body weight (F 2.4,70.3 = 0.7, P > 0.05, η 2 = 0.024), and no significant main effect for group (F 2,58 = 2.2, P > 0.05, η 2 = 0.070). The main effect for time was significant and large (F 1.2,70.3 = 183.8, P < 0.001, η 2 = 0.760). The percent of participants who met the weight loss goal of 10% in the PRO + EX, PRO, and CON + EX groups was 31.6%, 45.0%, and 50.0%, respectively, and did not differ significantly among groups (χ 2 (2) = 1.5, P > 0.05).

Body Composition. Body composition outcomes showed a similar pattern to weight loss. Both fat mass and %Fat showed no significant interaction (F 2.9,83.6 = 0.4, P > 0.05, η 2 = 0.014; and F 4,116 = 2.0, P > 0.05, η 2 = 0.066, respectively), and no significant group main effect (F 2,58 = 2.2, P > 0.05, η 2 = 0.070; and F 2,58 = 1.6, P > 0.05, η 2 = 0.051, respectively). Time main effects were significant for both variables (F 1.4,83.6 = 149.1, P < 0.001, η 2 = 0.720; and F 1.5,88.8 = 123.0, P < 0.001, η 2 = 0.680 for fat mass and relative adiposity, respectively). Results for lean mass (both whole-body and leg lean mass) were the same, with interaction effects being nonsignificant (F 4,116 = 2.0, P > 0.05, η 2 = 0.066 and F 3.3,95.7 = 0.7, P > 0.05, η 2 = 0.025 for whole-body and leg lean mass, respectively), and the main effects for group also nonsignificant (F 2,58 = 0.3, P > 0.05, η 2 = 0.038; and F 2,58 = 0.2, P > 0.05, η 2 = 0.007). Main effects for time were significant for both lean mass variables (F 2,116 = 7.7, P < 0.001, η 2 = 0.117; and F 1.7,95.7 = 7.4, P < 0.01, η 2 = 0.114).

Muscle Strength and Physical Function

Muscle Strength. There was a significant group–time interaction effect for IK-60 (F 4,116 = 5.1, P < 0.001, η 2 = 0.148), which was decomposed using Cohen's d effect sizes, focusing on changes across time within each group (Figure 2; Table 4). Comparatively, the net difference in IK-60 change over time represented a medium-to-large effect in favor of the two EX groups (d = 0.68 and d = 0.76 for CON + EX and PRO + EX, respectively) compared with the PRO group. The per-protocol analyses (n = 61) indicated that improvements in IK-60 from baseline to posttest were similar in PRO + EX and CON + EX (P > 0.05) and greater than PRO (both P < 0.05). The ITT analyses (n = 72) produced similar significance results (i.e., PRO + EX = CON + EX > PRO). Notably, neither the baseline nor the changes in weight or adiposity were related to changes in strength (all P > 0.05).

Figure 2.

A–D, Changes in muscle strength and physical function in response to the intervention. Numbers indicate effect size from baseline to posttest within group. IK-60, isokinetic knee torque at 60°·s−1; CHAIR, 30-s chair stand test; WALK, 6-min walk test.

Physical Function. Group–time interactions were significant for all physical function variables (UPGO F 3.1, 86.3 = 4.4, P < 0.01, η 2 = 0.136; CHAIR F 3.3, 95.1 = 11.2, P < 0.001, η 2 = 0.278; WALK F 3.2, 92.2 = 10.3, P < 0.001, η 2 = 0.262). Patterns of Cohen's d for UPGO revealed that in comparison with the PRO group, there was a net medium improvement for the two EX groups (d = −0.49 and −0.61 for CON + EX and PRO + EX, respectively). CHAIR scores were similar with the net difference being substantial when comparing the PRO to the improvements in both CON + EX (d = 1.62) and PRO + EX (d = 1.43). Finally, a similar pattern was seen for WALK with the EX groups exhibiting a moderate-to-large improvement compared with the PRO group (CON + EX d = 0.66 and PRO + EX d = 0.69, respectively). The per-protocol analyses (n = 61) indicated that improvements in UPGO, CHAIR, and WALK from baseline to posttest were similar in PRO + EX and CON + EX (P > 0.05) and greater than PRO (both P < 0.05). The ITT analyses (n = 72) produced similar significance results in all physical function outcomes (i.e., PRO + EX = CON + EX > PRO).

processing....