Single‐leg disuse decreases skeletal muscle strength, size, and power in uninjured adults: A systematic review and meta‐analysis

Abstract

We aimed to quantify declines from baseline in lower limb skeletal muscle size and strength of uninjured adults following single‐leg disuse. We searched EMBASE, Medline, CINAHL, and CCRCT up to 30 January 2022. Studies were included in the systematic review if they (1) recruited uninjured participants; (2) were an original experimental study; (3) employed a single‐leg disuse model; and (4) reported muscle strength, size, or power data following a period of single‐leg disuse for at least one group without a countermeasure. Studies were excluded if they (1) did not meet all inclusion criteria; (2) were not in English; (3) reported previously published muscle strength, size, or power data; or (4) could not be sourced from two different libraries, repeated online searches, and the authors. We used the Cochrane Risk of Bias Assessment Tool to assess risk of bias. We then performed random‐effects meta‐analyses on studies reporting measures of leg extension strength and extensor size. Our search revealed 6548 studies, and 86 were included in our systematic review. Data from 35 and 20 studies were then included in the meta‐analyses for measures of leg extensor strength and size, respectively (40 different studies). No meta‐analysis for muscle power was performed due to insufficient homogenous data. Effect sizes (Hedges' g_av) with 95% confidence intervals for leg extensor strength were all durations = −0.80 [−0.92, −0.68] (n = 429 participants; n = 68 aged 40 years or older; n ≥ 78 females); ≤7 days of disuse = −0.57 [−0.75, −0.40] (n = 151); >7 days and ≤14 days = −0.93 [−1.12, −0.74] (n = 206); and >14 days = −0.95 [−1.20, −0.70] (n = 72). Effect sizes for measures of leg extensor size were all durations = −0.41 [−0.51, −0.31] (n = 233; n = 32 aged 40 years or older; n ≥ 42 females); ≤7 days = −0.26 [−0.36, −0.16] (n = 84); >7 days and ≤14 days = −0.49 [−0.67, −0.30] (n = 102); and >14 days = −0.52 [−0.74, −0.30] (n = 47). Decreases in leg extensor strength (cast: −0.94 [−1.30, −0.59] (n = 73); brace: −0.90 [−1.18, −0.63] (n = 106)) and size (cast: −0.61[−0.87, −0.35] (n = 41); brace: (−0.48 [−1.04, 0.07] (n = 41)) following 14 days of disuse did not differ for cast and brace disuse models. Single‐leg disuse in adults resulted in a decline in leg extensor strength and size that reached a nadir beyond 14 days. Bracing and casting led to similar declines in leg extensor strength and size following 14 days of disuse. Studies including females and males and adults over 40 years of age are lacking.

1. Introduction

Maintaining skeletal muscle strength ¹ and size ² throughout life is linked to multiple positive health outcomes. ³ Unfortunately, periods of skeletal muscle disuse arising from illness, injury, and/or surgery result in a rapid decline in skeletal muscle strength and size. ⁴ These bouts of skeletal muscle disuse increase risk of metabolic disease and functional impairment leading to deleterious health consequences, especially in older adults. ⁵ , ⁶

To better understand the mechanisms driving losses of skeletal muscle strength and size in response to skeletal muscle disuse, researchers have employed a variety of experimental models such as bedrest, ⁷ single‐leg casting, ⁸ single‐leg bracing, ⁹ or single‐leg suspension. ¹⁰ Although bedrest elicits loss of skeletal muscle strength and size, bedrest studies require specialist facilities and medical oversight. Many researchers have instead relied upon single‐leg disuse models. Multiple original research studies^{8, 11–13} and reviews ¹⁴ , ¹⁵ have characterized the declines in skeletal muscle strength and size in response to single‐leg disuse. Only one meta‐analysis has quantified rates of decline in skeletal muscle size using 12 studies, nine of which involved complete bed rest. ¹⁶ Thus, a comprehensive up‐to‐date systematic review and meta‐analysis examining changes in skeletal strength and size following single‐leg disuse in uninjured adults would address this knowledge gap and serve as a valuable resource for those wishing to develop statistical power calculations in future work.

The effect of immobilization model on changes in skeletal muscle strength and size also remains underexplored. A recent systematic review reported greater declines in knee extensor strength in uninjured adults using unilateral lower limb suspension models compared with immobilization via cast or brace. ¹⁷ This observation is important because application of a cast often requires specialized healthcare providers and is more laborious to remove compared with bracing. A meta‐analysis quantifying the effect of single‐leg disuse model (e.g., casting vs. bracing) on declines in skeletal muscle strength and size in uninjured adults would build on this work ¹⁷ and help inform the choice of whether to use braces or casts in single‐leg immobilization research.

The primary purpose of this study was to conduct a systematic review and meta‐analysis of experimental studies that used a single‐leg disuse model in uninjured adults and reported changes in skeletal muscle strength, size, and power. Secondary purposes were to compare the effect of duration and single‐leg disuse model (i.e., casting vs bracing) on declines in skeletal muscle strength, size, and power.

2. Methods

The research question was established using the population, intervention, comparison, outcome, and setting criteria (see Sheet S1 in the supporting information). This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses checklist (see Sheet S3 in the supporting information). ¹⁸ Abstract/title screening and full‐text review were conducted using Covidence's systematic review software (Veritas Health Innovation, Australia). Intended methods were documented prior to the initial literature search, extraction, and analysis (Open Science Framework, osf.io/wdexu). An addendum was added on 30 January 2022 (osf.io/ut93s) to outline meta‐analytic and repeated search strategy plans.

2.1. Eligibility criteria

Studies were included if they met all the following inclusion criteria: (1) recruited non‐injured, human participants; (2) were an original, published experimental study; (3) employed a unilateral leg immobilization model; and (4) reported skeletal muscle size, strength, or power data in response to a period of single‐leg disuse for at least one group without a countermeasure (e.g., concomitant exercise training). Studies were excluded if they (1) did not meet all inclusion criteria; (2) were not in English; (3) reported previously published skeletal muscle strength, size, or power data; or (4) had full texts that could not be sourced from the Queen's University or University of Ottawa libraries, repeated online searches, or attempts to contact the authors.

2.2. Search strategy

We conducted an initial literature search in EMBASE and Medline on 14 December 2020, following protocol registration (osf.io/wdexu; osf.io/ut93s). A second, up‐to‐date search in EMBASE, Medline, Cochrane Central Register of Controlled Trials (CCRCT), and CINAHL took place on 20 January 2022 (osf.io/ut93s). No year limits were applied to the CINAHL and CCRCT searches; however, the English language limit was applied. The search strategies incorporated concepts of skeletal muscle disuse and skeletal muscle strength, size, and power. A complete list of keywords and MeSH terms for these main concepts were combined with ‘OR’, and the search strategies combined these lists with ‘AND’. English language and humans only limits were applied to both the Medline and EMBASE search strategies. The search strategies provided in Sheet S1 in the supporting information were developed and applied in collaboration with a Health Sciences Librarian at Queen's University (Sandra Halliday). Titles and abstracts were imported into Covidence from the database searches, and Covidence automatically removed duplicates. Two reviewers (N. P. and J. S.) also screened relevant previously published reviews citing single‐leg disuse ¹⁴ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ to identify eligible articles missed by our literature searches.

2.3. Selection process

Study selection followed a two‐step process and was independently completed by two reviewers (N. P. and J. S.). Both reviewers met to resolve conflicts, and a third reviewer (C. M.) provided a consensus vote for any remaining conflicts. First, titles and abstracts were screened to identify studies appearing to meet eligibility criteria. Second, full texts were downloaded for studies that passed title and abstract screening. Studies removed during full‐text screening were assigned a reason for exclusion (see Sheet S2 in the supporting information).

2.4. Risk of bias and quality of evidence assessments

Two reviewers (N. P. and J. S.) independently assessed the risk of bias in included studies using the Cochrane Collaboration Risk of Bias Assessment Tool. ²⁷ The risk of each source of bias was judged as either ‘high’, ‘low’, or ‘unclear’. Outcome‐specific judgements were completed for outcome assessor bias and attrition bias. Studies reporting an adequate methodology (e.g., blinding outcome assessors to protect against detection bias) in‐text or in a protocol registration document for protecting against a given source of bias were judged as having a ‘low’ risk of bias. Studies that did not report information regarding a given methodology were judged as having an ‘unclear’ risk of bias. Studies reporting ‘open‐label’ only in a registration document were judged as having a ‘high’ risk of performance and detection bias. Studies were judged as having a ‘high’ risk of reporting bias if they did not report publicly registering their trial. Study protocol documents, when available, informed judgements in cases of unclear or a lack of reporting in the published manuscript. Two reviewers (N. P. and C. M.) assessed quality of evidence using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) platform (https://gdt.gradepro.org/).

2.5. Data extraction

N. P. and J. S. independently performed data extraction of included studies. Measures of muscle strength, size, and power were extracted from original publications. Authors were not contacted for missing data. When studies reported previously published data, repeated data were not extracted. These two reviewers then compared extracted data and resolved any discrepancies. C. M. also randomly spot‐checked data extraction (N = 10 studies). Extracted outcome data and data items (e.g., participant characteristics and sources of funding) are provided in Sheets S4 and S5 in the supporting information. ²⁸ Sex and gender were interchanged (i.e., ‘woman’ with ‘female’ and ‘man’ with ‘male’) for data extraction purposes.

2.6. Data synthesis

Baseline (PRE), post‐disuse (POST), SD/SE, and per cent change were extracted either by recording values as reported in text or by using WebPlotDigitizer—a data extraction approach with high inter‐rater reliability and validity ²⁹ —when outcome data only appeared in figures. Blue‐filled cells in Sheet S5 in the supporting information denote data extracted using WebPlotDigitizer. We calculated and reported per cent changes because many studies report and use per cent changes. Per cent change was calculated for studies that only reported means at PRE and POST using the following equation:

$$\%\mathit{change}=\frac{\mathit{POST}-\mathit{PRE}}{\mathit{PRE}}\times 100$$

Meta‐analyses were planned for skeletal muscle strength, size, and power, but as outlined in our pre‐specified analysis plan (osf.io/wdexu), meta‐analyses were only conducted if there were ≥4 data points. Within‐ and between‐subject study designs were included in meta‐analyses (see Sheet S4 in the supporting information for study designs). Subgroup analyses were performed to determine if effect sizes differed according to length of disuse (≤7 days, >7 to ≤14 days, vs. >14 days) and model of disuse (casting vs. brace). The modal duration was selected for model‐specific analyses.

Studies were included in the meta‐analysis if they (1) reported a leg extensor strength or size outcome that was pre‐specified in osf.io/wdexu and (2) reported or had figures with extractable means and standard deviations or standard errors with sample sizes at PRE and POST. Studies only presenting data grouped by age or sex were entered as separate data points in meta‐analyses. Each meta‐analysis could only include a single outcome from a study sample since each contributing outcome needed to be from an independent sample. For studies with >1 measure of skeletal muscle strength or size, a single measure was included in the meta‐analysis. The strength measure used was selected in the following order of preference: isometric leg extension, isokinetic leg extension, and leg extension one repitition maximum. The muscle size measure used was selected in the following order of preference: quadriceps volume, quadriceps peak cross‐sectional area, vastus lateralis volume, and vastus lateralis peak cross‐sectional area as measured by magnetic resonance imagery or computed tomography. Studies not reporting these outcomes were excluded from the meta‐analyses.

We initially calculated Cohen's d _av for each outcome included in a meta‐analysis using the following formula ³⁰ :

$$\mathrm{Cohe}{\mathrm{n}}^{\text{'}}\mathrm{s}{\mathit{d}}_{\mathit{av}}=\frac{\mathit{POST}-\mathit{PRE}}{\sqrt{\frac{{\mathit{SD}}_{\mathit{POST}}^2+{\mathit{SD}}_{\mathit{PRE}}^2}{2}}}$$

This formula was developed for independent samples, ³¹ which is not ideal when studying pre‐to‐post changes within participants. However, d _av is a convenient solution when the correlation between PRE and POST measures, which is required in the most appropriate Cohen's d formulas, is unavailable. ³² Correlations between PRE and POST measures were unavailable in the current review as they were not identified and were seldom extractable (n = 3 ¹² , ³³ , ³⁴ ). Negative effect sizes with larger magnitudes indicated a larger loss of skeletal muscle strength or size because effect sizes were calculated by subtracting the mean POST from the mean PRE value.

Given Cohen's d is biased upwards for sample sizes less than 20, ³¹ and all outcomes from included studies had sample sizes less than 20, we converted Cohen's d _av values to Hedges' g _av values using the following formula:

$${\text{Hedges}}^{\text{'}}{\mathit{g}}_{\mathit{av}}={\mathit{d}}_{\mathit{av}}*\left(1-\left(\frac{3}{4*\mathit{n}-1}-1\right)\right)$$

The sample size at PRE and POST, n, was always identical for a given outcome of a given study. The precision of Hedges' g _av effect size estimates was then determined by calculating the standard error (SE_g) for each Hedges' g value using the following equation ³² :

$${\mathit{SE}}_{\mathit{g}}=\sqrt{\left(\frac{\mathit{n}+\mathit{n}}{2*\mathit{n}}\right)+\left(\frac{{\mathit{g}}_{\mathit{av}}^2}{2\left(\mathit{n}+\mathit{n}\right)}\right)}$$

After Hedges' g _av were determined, random‐effects model meta‐analyses were conducted using Meta‐Essentials software version 1.4. ³⁵ This software was also used to create forest and funnel plots. Forest plots present Hedges' g _av with 95% confidence intervals and prediction intervals (i.e. ‘the range in which the point estimate of 95% of future studies will fall, assuming that true effect sizes are normally distributed through the domain’ ³⁶ ). For studies including more than one time point, ³⁷ , ³⁸ , ³⁹ the data point contributing to the modal duration was selected. Heterogeneity was tested using the I ² test and significance was set at P < 0.05. I ² values of 25%, 50%, and 75% were considered low, moderate, and high, respectively. ⁴⁰ Funnel plots and Egger's tests were used to determine presence of publication bias. ⁴¹ Adjusted combined effect sizes in funnel plots represent results after applying the trim‐and‐fill method. ⁴²

3. Results

3.1. Systematic review

3.1.1. Study selection

Figure 1 presents a flow diagram of the study selection process. The first literature search retrieved 6376 studies, and Covidence removed 625 duplicates. The updated literature search retrieved 1509 studies, and Covidence removed 712 duplicates. Six thousand five hundred forty‐eight studies entered title and abstract screening, and 6299 were deemed irrelevant and were subsequently excluded. Full texts of 249 studies were downloaded, and 163 studies were excluded (reasons provided in Figure 1 and Sheet S2 in the supporting information), leaving 86 included studies. This number includes one study ⁴³ that was added after screening relevant and previously published reviews mentioning single‐leg disuse. References for all 86 included studies in the systematic review are provided at the end of this manuscript (see Appendix A).

Figure 1.

Flow diagram of the study selection process. Abbreviations: CCRCT, Cochrane Central Register of Controlled Trials; CINAHL, Cumulative Index to Nursing and Allied Health Literature; CSA, cross‐sectional area

3.1.2. Study characteristics, safety, and risk of bias

Sheets S4 and S5 in the supporting information present characteristics of the 86 included studies and their participants, and Sheet S5 in the supporting information presents results of individual studies. Seventy‐six (88%) and 48 (55%) of included studies were published since 2000 and 2010, respectively. Fifty‐four studies (63%) used a non‐cast immobilization model. Forty‐four (51%) of the included studies reported strength and size outcomes; 18 (21%) reported strength outcomes only; 15 (17%) reported size outcomes only; 5 (6%) reported strength, size, and power outcomes; 3 (3%) reported strength and power outcomes; and 1 (1%) reported size and power outcomes. Thirty‐one studies (36%) included females. Of these, six studies (7%) included only females, ⁹ , ³⁸ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ and only one study including males and females presented sex‐disaggregated data. ³⁷ Eleven studies (13%) included individuals >40 years old. ¹¹ , ¹² , ³⁴ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵

Twenty (23%) studies reported performing a sample size calculation. Forty‐two (49%) and 38 (44%) studies reported compliance and safety measures related to disuse, respectively. Additionally, most studies were judged to have an unclear or high risk of selection, performance, detection, and reporting bias (Sheet S6 in the supporting information).

3.1.3. Average per cent changes in leg extensor strength, size, and power

Sheet S7 in the supporting information displays results from studies reporting declines in any measure of knee extensor strength and whole quadriceps muscle size across all durations of disuse. The decline (mean ± SD) in knee extension strength was −13 ± 7% (range: −28 to −6%; N = 7 studies), and the decline in whole quadriceps muscle size was −5 ± 2% (−7% to −2%; N = 4) at 7 days of disuse. At 14 days of disuse, the decline in knee extension strength was −23 ± 6% (−36 to −10%; N = 24), and the decline in whole quadriceps muscle size was −8 ± 3% (−15 to −5%; N = 12). For disuse durations less than 7 days, the decline in knee extension strength was −10 ± 6% (−24 to −2%, N = 6), and the decline in whole quadriceps muscle size was −3 ± 1% (−4 to −2%, N = 5). From 7 days and ≤14 days, the decline in knee extension strength was −21 ± 8% (−42 to −6, N = 33), and the decline in whole quadriceps muscle size was −8 ± 3% (−15 to −5, N = 17). For studies of 14 days of disuse and longer, the decline in knee extension strength was −25 ± 11% (−52 to −16, N = 15) and the decline in whole quadriceps muscle size was −12 ± 5% (−21 to −7, N = 8). Sheet S5 in the supporting information also contains data from nine separate studies that examined changes in skeletal muscle power. The most common measure was knee extensor isokinetic power, and it declined by −11 ± 1 (−12 to −9) at 0.53 rad/s, −5 ± 3 (−10 to −2) at 1.05 rad/s, and −6 ± 9 (−16 to 2) at 2.09 rad/s in young, old, trained, and untrained adults after 7 days of disuse. ⁵⁰ , ⁵⁶

3.2. Meta‐analyses

3.2.1. Study characteristics, safety, and risk of bias

Of the 86 studies included in the present review, 40 were then included in meta‐analyses (N = 35 studies reported knee extensor data; N = 20 studies reported knee extensor size; N = 15 studies reported both). Sheet S8 in the supporting information provides the list of studies excluded from a meta‐analysis and their reasons for exclusion. The strength meta‐analysis included 35 separate studies, and 14 (40%) of which included females. Of these, three studies (9%) included females only, ⁹ , ³⁸ , ⁴⁵ and one study (3%) including males and females presented sex‐disaggregated data. ³⁷ Six studies (17%) included individuals 40 years of age or older with four of these studies (11%) presenting age‐disaggregated data for older and younger adults. ¹² , ³⁴ , ⁵⁰ , ⁵³ The leg extensor size meta‐analysis included 20 separate studies, and eight (40%) of which included females. Of these studies, only a single study (5%) included females only, ⁹ and only one study (5%) including males and females presented sex‐disaggregated data. ³⁷ Three studies (15%) included individuals 40 years of age or older with a single study (5%) presenting age‐disaggregated data for older and younger adults. ¹² Although nine studies reported measures of power, three studies measured isokinetic muscle power and presented five data points total. ⁵⁰ , ⁵⁶ , ⁵⁷ Of those studies, two studies (four data points) measured leg extension power at 7 days of disuse with ≤2.09 rad/s, and one study measured after 14 days of disuse with 3.14 rad/s. Furthermore, all three studies also presented data from different populations (e.g., trained vs. untrained, old vs. young). This meant no skeletal muscle power outcome had four comparable data points or were sufficiently homogenous to warrant a meaningful meta‐analysis for muscle power; a meta‐analysis on muscle power was therefore not performed.

Eleven (31%) and seven (35%) studies reported performing a sample size calculation in the knee extensor strength and size meta‐analyses, respectively. Eighteen (51%) and 9 (45%) studies in the strength and size meta‐analyses reported a compliance measure for disuse, respectively. Twenty (57%) and 12 (60%) studies in the strength and size meta‐analyses reported a safety measure for disuse, respectively. Additionally, most studies were judged to have an unclear or high risk of selection, performance, detection, and reporting bias (Sheet S6 in the supporting information).

3.2.2. Publication bias

I ² statistics were 0% for muscle strength and size meta‐analyses (except strength measures across all durations: 10.1% when trim and fill method imputed six studies), and this indicated homogeneity across all comparisons. Although visual inspection of funnel plots and Egger's P‐values (strength: P = 0.03; size: P = 0.16) revealed potential publication bias when all measures of strength and all measures of size were combined across all durations of disuse (see Sheet S10 in the supporting information), the observation that skeletal muscle size and strength never increased and is not expected to increase in response to disuse strongly questions the presence of true publication bias omitting studies with positive effect sizes.

3.2.3. Leg extensor strength, size, and power

Figure 2 presents the forest plot for measures of leg extensor strength across all durations (weighted effect size with 95% confidence interval = −0.80 [−0.92, −0.68], P < 0.001, N = 35 studies, n = 429 participants). Figure 3 presents the forest plot for measures of leg extensor size across all durations (−0.41 [−0.51, −0.31], P < 0.001, N = 20, n = 233). Figure 4 and Sheet S9 in the supporting information present subgroup meta‐analyses and their weighted effects sizes for measures of leg extensor strength (≤7 days (−0.57 [−0.75, −0.40], N = 9, n = 151), >7 days and ≤14 days (−0.93 [−1.12, −0.74], N = 17, n = 206), and >14 days (−0.95 [−1.20, −0.70, N = 9, n = 72) and leg extensor muscle size (≤7 days (−0.26 [−0.36, −0.16], N = 7, n = 84), >7 days and ≤14 days (−0.49 [−0.67, −0.30], N = 8, n = 102), and >14 days of disuse (−0.52 [−0.74, −0.30], N = 6, n = 47). Although muscle power data was not meta‐analysed, Sheet S5 in the supporting information includes data from the nine studies that measured muscle power.

Figure 2.

Forest plot of a random effects meta‐analysis comparing changes in measures of leg extensor strength following single‐leg disuse. Because effect sizes were calculated as baseline minus post‐disuse values, negative Hedge's g values with larger magnitudes reflect larger declines in muscle strength. The red circle represents a combined effect size with 95% confidence intervals. Outcomes reported in N, kg, and nm were categorized as MVC, 1RM, and peak torque, respectively (unless reported otherwise). Table S5 in the supporting information includes all extracted outcome data; n = 9 studies ≤7 days, n = 17 studies >7 to ≤14 days, n = 9 studies >14 days; N = 35 unique studies. (a) trained subgroup; (b) untrained subgroup; (c) old subgroup; (d) young subgroup; (e) male subgroup; (f) female subgroup; * subgroups reported in the same publication. AJCN, American Journal of Clinical Nutrition; CI, confidence interval; MSSE, Medicine & Science in Sports & Exercise; MVC, maximum voluntary contraction; NR, not reported; 1RM, one‐repetition maximum

Figure 3.

Forest plot of a random effects meta‐analysis comparing changes in measures of leg extensor size following single‐leg disuse. Because effect sizes were calculated as baseline minus post‐disuse values, negative Hedge's g values with larger magnitudes reflect larger declines in muscle size. The red circle represents a combined effect size with 95% confidence intervals. N = 20 unique studies. Sheet S5 in the supporting information includes all extracted outcome data; (a) female subgroup; (b) male subgroup; (c) old subgroup; (d) young subgroup; * subgroups reported in the same publication. AJCN, American Journal of Clinical Nutrition; AJPEM, American Journal of Physiology, Endocrinology, and Metabolism; CI, confidence interval; CSA, cross‐sectional area; NR, not reported; VL, vastus lateralis

Figure 4.

Subgroup analyses for changes in measures of knee extensor strength and size following single‐leg disuse. Red circles represent combined effect sizes with 95% confidence intervals. Sheet S9 in the supporting information provides subgroup analyses with individual studies. Knee extensor strength: N = 35 separate studies; n = 429 observations; I ² = 0% for all. ANOVA for subgroup differences (between/model): Sum of squares = 5.98, df = 2, P = 0.05. Knee extensor strength (cast vs. brace following 14 days of disuse): N = 14 separate studies; n = 179 observations; I ² = 0% for all. ANOVA for subgroup differences (between/model): Sum of squares = 0.03, df = 1, P = 0.86. Knee extensor size: N = 20 separate studies; n = 233 observations; I ² = 0% for all. ANOVA for subgroup differences (between/model): Sum of squares = 1.56, df = 2, P = 0.46. Knee extensor size (cast vs. brace following 14 days of disuse): N = 6 separate studies; n = 82 observations; I ² = 0% for all. ANOVA for subgroup differences (between/model): Sum of squares = 0.17, df = 1, P = 0.68

3.2.4. Leg extensor strength and size by disuse model

The most common duration of disuse was 14 days. Figure 4 and Sheet S9 in the supporting information present exploratory subgroup meta‐analyses and their weighted effect sizes for leg extensor strength using casts (−0.94 [−1.30, −0.59], N = 7 studies, n = 73 participants) and braces (−0.90 [−1.18, −0.63], N = 7, n = 106) and leg extensor size using casts (−0.61 [−0.87, −0.35], N = 3, n = 41) and braces (−0.48 [−1.04, 0.07], N = 3, n = 41).

4. Discussion

This systematic review and meta‐analysis quantified changes in skeletal muscle strength and size following single‐leg disuse in uninjured adults. The primary finding was that single‐leg disuse results in a significant decline in leg extensor muscle strength and size with the magnitude of these declines plateauing after 14 days of disuse. There were too few data points to perform a meaningful meta‐analysis on changes in muscle power following single‐leg disuse. We also provide evidence that the type of single‐leg disuse model (i.e., brace vs. cast) does not appear to modulate losses in leg extensor muscle strength and size following 14 days of disuse. Finally, we uncovered a lack of studies that included adults over the age of 40 years and those that interrogated the impact of sex on changes in skeletal muscle strength, size, and power in response to single‐leg disuse.

Consistent with several original research reports, ⁸ , ⁹ , ⁵¹ , ⁵⁸ narrative reviews, ¹⁴ , ²⁴ , ⁵⁹ and a systematic review, ¹⁷ our meta‐analyses demonstrate that single‐leg disuse results in a decline in skeletal muscle strength and size. We also show that durations of disuse longer than 7 days led to further decreases in skeletal muscle strength and size that reached a nadir beyond 14 days. The declines in skeletal muscle strength were greater compared to those of skeletal muscle size (effect sizes of −0.80 [−0.94, −0.67] and −0.41 [−0.52, −0.30] for leg extensor strength and size, respectively), further highlighting the disconnect between changes in skeletal muscle size and strength with unloading. ¹⁷ Our finding of greater declines in strength compared with size in response to single‐leg disuse is particularly relevant to older adults as skeletal muscle strength associates with numerous clinically relevant outcomes in this population. ⁶⁰ , ⁶¹ However, only nine and five studies in our systematic review and meta‐analyses, respectively, included adults aged 65 years or older. Only two studies in the systematic review and one study in the meta‐analysis included adults with average ages between 40 to 50 years old. ⁴⁸ , ⁵⁴ This lack of studies prevented meaningful exploratory meta‐analyses sub‐grouped by age. The dearth of single‐leg disuse studies in older and middle‐aged adults may be partially explained by observations that older adults display impaired regenerative capacity compared with their younger counterparts ⁶² , ⁶³ and that middle‐aged adults often have occupational commitments which limit their participation in muscle disuse studies. Given the importance of mitigating declines in strength during healthy aging ⁶⁰ , ⁶¹ and the rapidly growing number of older adults worldwide, ⁶⁴ examining how healthy older and middle‐aged adults respond to single‐leg disuse therefore represents an important focus of future work.

A previous systematic review compared the effects of fixed angle (casting and bracing) versus free joint angle (unilateral lower limb suspension) models of immobilization on changes in skeletal muscle strength and size in response to skeletal muscle disuse. ¹⁷ The authors concluded that fixed immobilization methods resulted in greater skeletal muscle strength losses, but similar declines in skeletal muscle size compared with free‐joint angle approaches. We build on these findings by showing that changes in skeletal muscle strength and size were not different between fixed cast and fixed brace models of single‐leg disuse. Our data also complement a recent multicentre randomized controlled trial in adults with ankle fractures that showed no difference in Olerud Molander ankle scores – a functional rating scale – between ankle casting and ankle bracing. ⁶⁵ These observations ¹⁷ , ⁶⁵ and our data together suggest fixed cast and bracing models could be used interchangeably in disuse studies. However, the interpretation that cast and brace models could be used interchangeably in single‐leg disuse studies including uninjured adults should be viewed cautiously as our exploratory meta‐analyses comparing these models is limited to a 14‐day time point with only four data points for each disuse model in the analysis for leg extensor size. Studies directly comparing the effect of cast versus brace on the magnitude of muscle strength/size loss in response to single‐leg disuse will help draw firmer conclusions.

The range of skeletal muscle strength and size responses to single‐leg disuse is another interesting observation of our study. For instance, reductions in knee extensor strength following 14 days of single‐leg immobilization ranged from −10% ³⁷ to −36%, ⁶⁶ and changes in quadriceps size ranged from −5% ¹² , ³⁹ to −14.5%. ⁹ Examining the reasons underlying the heterogeneity in response to single‐leg disuse falls beyond the purview of the present report; these reasons have been discussed elsewhere. ⁶⁷ We speculate that differences in measurement techniques, and to a lesser degree, genetic, and/or transcriptional factors contribute to this heterogeneity. Exploring the differential impact of these sources of heterogeneity on declines in skeletal muscle strength and size in response to muscle disuse requires additional experimental attention. The wide confidence intervals stemming from small sample sizes within a given study led to downgrading the certainty of evidence in each of our GRADE assessments (Sheet S11 in the supporting information). Furthermore, we found an unclear or high risk of bias across several domains of bias (Sheet S6 in the supporting information), and this suggests our weighted effect sizes could be imprecise because studies inadequately protecting against bias can inflate effect sizes. ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² Future work would therefore benefit from incorporating bias‐reducing practices as advocated for previously. ⁷³ , ⁷⁴ , ⁷⁵

To our knowledge, our study is the largest to meta‐analyse declines in both skeletal muscle strength and size following single‐leg disuse in uninjured adults. It therefore provides a more precise estimate of the true effect of single‐leg disuse on these outcomes and novel data for a priori statistical power calculations in research involving single‐leg disuse models. The need for such data is highlighted by the fact that only 23% of studies included in our systematic review reported using an a priori statistical power calculation. Moreover, by excluding studies that assessed the effects of single‐leg disuse due to injury on skeletal muscle strength and size, we avoided known confounding conditions such as hyperinflammation ⁷⁶ and hypercortisolaemia. ⁷⁷ Excluding these studies may limit the translatability of our findings to clinical settings; however, we propose that our data are highly relevant to those laboratories using ‘uncomplicated’ disuse models aimed at developing nutritional, exercise, and/or pharmacological countermeasures to the negative impacts of skeletal muscle disuse. Future meta‐analyses could complement our findings by incorporating single‐leg disuse models involving injured participants to ascertain whether similar declines in muscle strength and size occur between injury and non‐injury scenarios of single‐leg muscle disuse. For example, a recent report demonstrated that the critically ill experienced significantly greater rates of lower limb skeletal muscle loss compared with healthy individuals undergoing muscle disuse. ¹⁶ This result highlights how our findings in uninjured adults may not necessarily recapitulate clinical scenarios of disuse.

Despite including 86 studies in the systematic review and 40 in the meta‐analyses, we did not perform sex‐based analyses as insufficient studies included females. The lack of single‐leg disuse studies that include females is difficult to justify. Females are more susceptible to injury and subsequent limb disuse than males. ⁷⁸ Yet few studies examine how disuse affects skeletal muscle in females (36% of studies in the systematic review, n = 31) or in a sexually dimorphic manner (1%, n = 1). Single‐leg disuse studies involving female participants and disaggregating data by sex therefore represents a clear priority for future work.

In conclusion, this systematic review and meta‐analysis quantified the effect of single‐leg muscle disuse on skeletal muscle strength and size in uninjured adults. We also provide evidence that fixed casting and bracing models result in similar declines in leg extensor strength and size. Finally, we highlight a lack of data generated in middle‐aged adults, older adults, and females undergoing single‐leg disuse. Studies examining sex‐specific and age‐specific adaptations in skeletal muscle strength, size, and power in response to single‐leg disuse are now needed.

Funding

This work was funded by a research initiation grant from Queen's University to Chris McGlory. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of interest

All authors declare no conflict of interest.

Supporting information

Data S1. Supporting Information

APPENDIX A.

⁹ , ¹¹ , ¹² , ¹³ , ¹⁵ , ³³ , ³⁴ , ³⁷ , ³⁹ , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁶⁶ , ⁶⁷ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ , ⁸⁷ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ , ⁹⁵ , ⁹⁶ , ⁹⁷ , ⁹⁸ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ , ¹⁰⁶ , ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹² , ¹¹³ , ¹¹⁴ , ¹¹⁵ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸ , ¹¹⁹ , ^{120 *}

*These references also appear in the reference list in the manuscript and Sheet S12 in the supporting information.

Preobrazenski N, Seigel J, Halliday S, Janssen I, McGlory C. (2023) Single‐leg disuse decreases skeletal muscle strength, size, and power in uninjured adults: A systematic review and meta‐analysis, Journal of Cachexia, Sarcopenia and Muscle, 14, 684–696, 10.1002/jcsm.13201