Page 25 - 《运动与健康科学》(英文)2024年第2期

TaggedAPTARAEndResistance training volume effects on menopause 147
Comparators (CG participants who did not receive the RT After evaluation of these domains, the overall bias was consid-
intervention), Outcome (body adiposity, metabolic risk, and ered through the risk-of-bias judgment based on low/high/some
concerns. The quality assessments of both reviewers (PRPN and
inflammation assessments), and Study design (RCTs).TaggedAPTARAEnd
TaggedAPTARAPInitially, duplicate records were excluded. Subsequently, the PCS) were compared, and disagreements in the scores were
records were retrieved for screening by title and abstract. The resolved by discussion (unanimous agreement).TaggedAPTARAEnd
articles were included in the review if they met the following
criteria: (a) RCT study; (b) postmenopausal females, character- TaggedAPTARAH22.4. Certainty of the evidence: GRADE approachTaggedAPTARAEnd
ized by amenorrhea (at least 1 year) and aged 45 years, or
TaggedAPTARAPTwo authors (MAdSC and PRPN) independently performed
older females with age 60 years old; (c) intervention group
the Grading of Recommendations, Assessment, Development,
with supervised land based RT for the major muscle groups
and Evaluation (GRADE) approach, which was followed to
with specified intensity, volume, and exercises; (d) comparator
assess the certainty of the evidence supporting the effects of
CG with no exercise and/or placebo interventions (i.e., low 58
RT on each outcome. According to the GRADE approach, 5
energy expenditure that did not affect the outcomes, such as
factors reduce the certainty of the evidence: (a) risk of bias or
stretching exercises); and (e) measurements (literature estab-
limitations in the detailed design and implementation, (b)
lished measurement methods) from baseline to the last available
unexplained heterogeneity or inconsistency of results, (c) indi-
follow-up of body adiposity, serum/plasma inflammatory
rectness of evidence, (d) imprecision of results, and (e) high
profile, and/or serum/plasma metabolic profile. Studies were
probability of publication bias. The certainty of evidence was
excluded according to the following criteria: females that
rated as high, moderate, low, or very low. Disagreements in
received any type of hormonal therapy or phytoestrogens; or
the GRADE approach were resolved by discussion.TaggedAPTARAEnd
were engaged any nutrition strategies (caloric restriction) for
weight loss; or were sufficiently active females, according to
TaggedAPTARAH22.5. Data syntheses and analysesTaggedAPTARAEnd
the World Health Organization recommendations involving the
practice of physical activity (e.g., aerobic or resistance exercises TaggedAPTARAPThe meta-analysis was conducted using Review Manager
that could affect the outcomes) and were engaged experimental Software (Version 5.4.; RevMan (Computer program), the
intervention using no structured physical exercise other than Cochrane Collaboration, London, UK). RevMan was used to
RT (e.g., aerobic exercises). The study selection agreement calculate the effect size of the RT intervention on each
between MAdSC and PRPN presented a Cohen’s k outcome (Tables 3 and 4) separately: body adiposity (total
result = 0.865, p < 0.001. Any study selection disagreements body adiposity and abdominal adiposity), and serum/plasma
were discussed with a third author (PCS). The third author was inflammatory (CRP) and metabolic profiles (TC, LDL-c,
HDL-c, TG, and fasting glucose). The variation (pre- minus
responsible for the tiebreaking decision.TaggedAPTARAEnd
post-intervention) from all included studies was used to calcu-
late the standardized mean difference and 95% confidence
TaggedAPTARAH22.3. Data extraction and quality assessments of each studyTaggedAPTARAEnd
interval (95%CI), and these were conducted using the DerSi-
TaggedAPTARAPOne author (PRPN) extracted the following data from each monian-Laird random-effects inverse variance model for all
study for analysis: author/year, number of participants within outcomes. Weighted percentages were based on the sample
each group, baseline participant characteristics, intervention sizes of the respective studies. Statistical significance was
details, and pre- and post-data from all outcomes (Tables 1 assumed as p < 0.05 in a Z-test analysis to examine whether
and 2). For studies containing multiple intervention arms vs. the effect size differed significantly from 0.TaggedAPTARAEnd
2
CGs, only data from RT groups were extracted. After the data TaggedAPTARAPStudy heterogeneity was evaluated using the I statistic and
2
extraction, the authors (AAdO, BdFC, GCS, and LMVS) inde- Cochrane’s Q. Values of I higher than 50% and 75% were
pendently confirmed the precision of the extracted data. considered moderate and high heterogeneity, with a threshold
Contact with corresponding authors was performed to clarify p 0.1. For Cochrane’s Q, significant heterogeneity exists
data or obtain missing information. Each outcome for body when the Q value exceeds the degrees of freedom of the esti-
adiposity (e.g., total body adiposity and abdominal adiposity) mate. Moreover, publication bias was tested visually using a
and serum/plasma inflammatory (e.g., CRP, IL-6, and tumor funnel plot when a sufficiently large sample of studies (i.e.,
necrosis factor-a) and metabolic profile (e.g., total cholesterol (TC), 10 study groups) was available for the RT vs. CG compar-
high-density lipoprotein cholesterol (HDL-c), low-density ison. Sensitivity analyses were performed by excluding 1 trial
lipoprotein cholesterol (LDL-c), triglycerides (TG), and at a time, according to the risk of bias, to test the robustness of
fasting glucose) were extracted separately for later analysis the pooled results. Effect sizes (Hedges’ g) were calculated,
with values of 0.000.19 considered trivial, >0.190.49
(Tables 3 and 4).TaggedAPTARAEnd
TaggedAPTARAPThe quality of the included studies was assessed using the small, >0.490.79 moderate, and >0.79 large.TaggedAPTARAEnd
revised tool for assessing the risk of bias in randomized trials. 57 TaggedAPTARAPThe RT studies were coded according to the total training
The assessment was performed according to the evaluation of volume (dosage), calculated as: number of exercises £ number
certain domains: (a) bias arising from the randomization of sets per exercise £ number of reps per exercise £ number
process, (b) bias due to deviations from intended interventions, of exercise sessions per week £ weeks of intervention. 59
(c) bias due to missing outcome data, (d) bias in measurement Regarding RT prescription and data imputation, if a range of
of the outcome, and (e) bias in selection of the reported result. values was used in the studies, for example 23 days per